Compare commits

...
198 Commits
Author SHA1 Message Date
antirez 987a47b823 Redis 2.8.10. 2014-06-05 11:02:23 +02:00
antirez 430bec3461 Don't process min-slaves-to-write for slaves.
Replication is totally broken when a slave has this option, since it
stops accepting updates from masters.

This fixes issue #1434.
2014-06-05 10:50:31 +02:00
antirez 3a30be51aa Tests for min-slaves-* feature. 2014-06-05 10:50:31 +02:00
antirez 42504169c7 Fixed dbuf variable scope in luaRedisGenericCommand().
I'm not sure if while the visibility is the inner block, the fact we
point to 'dbuf' is a problem or not, probably the stack var isx
guaranteed to live until the function returns. However obvious code is
better anyway.
2014-06-04 18:58:58 +02:00
antirez 26359d570b Regression test for issue #1118. 2014-06-04 18:51:31 +02:00
antirez 768994b683 Scripting: better Lua number -> string conversion in luaRedisGenericCommand().
The lua_to*string() family of functions use a non optimal format
specifier when converting integers to strings. This has both the problem
of the number being converted in exponential notation, which we don't
use as a Redis return value when floating point numbers are involed,
and, moreover, there is a loss of precision since the default format
specifier is not able to represent numbers that must be represented
exactly in the IEEE 754 number mantissa.

The new code handles it as a special case using a saner conversion.

This fixes issue #1118.
2014-06-04 18:45:49 +02:00
antirez 1c0c42e668 More trailing spaces in sentinel.c removed. 2014-05-28 15:46:14 +02:00
Matt Stancliff 75266018ee Disable recursive watchdog signal handler
If we are in the signal handler, we don't want to handle
the signal again.  In extreme cases, this can cause a stack overflow
and segfault Redis.

Fixes #1771
2014-05-26 17:53:40 +02:00
antirez edc41baca5 Sentinel example config: explain you don't need to specify slaves. 2014-05-26 10:18:07 +02:00
antirez 1a14eba2d5 Test: fixed scripting.tcl test false positive. 2014-05-22 15:46:06 +02:00
antirez f482349707 Process events with processEventsWhileBlocked() when blocked.
When we are blocked and a few events a processed from time to time, it
is smarter to call the event handler a few times in order to handle the
accept, read, write, close cycle of a client in a single pass, otherwise
there is too much latency added for clients to receive a reply while the
server is busy in some way (for example during the DB loading).
2014-05-22 15:46:02 +02:00
antirez f3d3c606f1 Accept multiple clients per iteration.
When the listening sockets readable event is fired, we have the chance
to accept multiple clients instead of accepting a single one. This makes
Redis more responsive when there is a mass-connect event (for example
after the server startup), and in workloads where a connect-disconnect
pattern is used often, so that multiple clients are waiting to be
accepted continuously.

As a side effect, this commit makes the LOADING, BUSY, and similar
errors much faster to deliver to the client, making Redis more
responsive when there is to return errors to inform the clients that the
server is blocked in an not interruptible operation.
2014-05-22 15:45:58 +02:00
antirez 4113473997 While ANET_ERR is -1, check syscall retval for -1 itself. 2014-05-22 15:45:54 +02:00
antirez e8669ca8a7 Regression test for issue #1764. 2014-05-20 16:20:45 +02:00
michael-grunder d491a479bc Fix LUA_OBJCACHE segfault.
When scanning the argument list inside of a redis.call() invocation
for pre-cached values, there was no check being done that the
argument we were on was in fact within the bounds of the cache size.

So if a redis.call() command was ever executed with more than 32
arguments (current cache size #define setting) redis-server could
segfault.
2014-05-20 16:20:38 +02:00
antirez d575f7a147 Remove trailing spaces from scripting.c 2014-05-20 16:11:28 +02:00
antirez 6d2fddd2a3 Remove trailing spaces from sentinel.c. 2014-05-20 14:22:34 +02:00
antirez ad935769ae HyperLogLog regression test for issue #1762. 2014-05-19 15:45:24 +02:00
Mike Trinkala 43278af63e Correct the HyperLogLog stale cache flag to prevent unnecessary computations.
Set the MSB as documented.
2014-05-19 15:45:24 +02:00
antirez 3203e2dcdd Fix 2.8 backport of fastscript branch.
No REDIS_ENCODING_EMBSTR in 2.8 internals.
2014-05-14 18:39:42 +02:00
Akos Vandra 6e478ec890 Fixed possible buffer overflow bug if RDB file is corrupted.
(Note: commit message modified by @antirez for clarity).
2014-05-12 11:49:16 +02:00
Akos Vandra c0eaac9572 fixed possible buffer overflow error 2014-05-12 11:19:36 +02:00
antirez 23011dea1f Sentinel: Add "dir /tmp" directive in example sentinel.conf. 2014-05-12 10:46:56 +02:00
antirez 6ae4c1fcbc DEBUG POPULATE: call dictExpand() to avoid useless rehashing. 2014-05-09 15:05:31 +02:00
antirez 13d8b2b096 Sentinel: log when a failover will be attempted again.
When a Sentinel performs a failover (successful or not), or when a
Sentinel votes for a different Sentinel trying to start a failover, it
sets a min delay before it will try to get elected for a failover.

While not strictly needed, because if multiple Sentinels will try
to failover the same master at the same time, only one configuration
will eventually win, this serialization is practically very useful.
Normal failovers are cleaner: one Sentinel starts to failover, the
others update their config when the Sentinel performing the failover
is able to get the selected slave to move from the role of slave to the
one of master.

However currently this timeout was implicit, so users could see
Sentinels not reacting, after a failed failover, for some time, without
giving any feedback in the logs to the poor sysadmin waiting for clues.

This commit makes Sentinels more verbose about the delay: when a master
is down and a failover attempt is not performed because the delay has
still not elaped, something like that will be logged:

    Next failover delay: I will not start a failover
    before Thu May  8 16:48:59 2014
2014-05-08 16:49:45 +02:00
antirez 909d1883e7 Sentinel: generate +config-update-from event when a new config is received.
This event makes clear, before the switch-master event is generated,
that a Sentinel received a configuration update from another Sentinel.
2014-05-08 16:49:45 +02:00
antirez 80b96bc4a7 Scripting test: check that Lua can call commands rewirting argv.
SPOP, tested in the new test, is among the commands rewritng the
client->argv argument vector (it gets rewritten as SREM) for command
replication purposes.

Because of recent optimizations to client->argv caching in the context
of the Lua internal Redis client, it is important to test for SPOP to be
callable from Lua without bad effects to the other commands.
2014-05-07 16:17:11 +02:00
antirez 093eab56de Test: handle new osx 'leaks' error.
Sometimes the process is still there but no longer in a state that can
be checked (after being killed). This used to happen after a call to
SHUTDOWN NOSAVE in the scripting unit, causing a false positive.
2014-05-07 16:17:11 +02:00
antirez 32971f6251 Scripting: objects caching for Lua c->argv creation.
Reusing small objects when possible is a major speedup under certain
conditions, since it is able to avoid the malloc/free pattern that
otherwise is performed for every argument in the client command vector.
2014-05-07 16:17:11 +02:00
antirez 87516f1a4d Scripting: Use faster API for Lua client c->argv creation.
Replace the three calls to Lua API lua_tostring, lua_lua_strlen,
and lua_isstring, with a single call to lua_tolstring.

~ 5% consistent speed gain measured.
2014-05-07 16:17:11 +02:00
antirez 1b5524f697 Scripting: don't call lua_gc() after Lua script run.
Calling lua_gc() after every script execution is too expensive, and
apparently does not make the execution smoother: the same peak latency
was measured before and after the commit.

This change accounts for scripts execution speedup in the order of 10%.
2014-05-07 16:17:11 +02:00
antirez 8237d88f48 Scripting: cache argv in luaRedisGenericCommand().
~ 4% consistently measured speed improvement.
2014-05-07 16:17:11 +02:00
antirez 6938a3d58f Fixed missing c->bufpos reset in luaRedisGenericCommand().
Bug introduced when adding a fast path to avoid copying the reply buffer
for small replies that fit into the client static buffer.
2014-05-07 16:17:11 +02:00
antirez 08ea98c9e9 Scripting: replace tolower() with faster code in evalGenericCommand().
The function showed up consuming a non trivial amount of time in the
profiler output. After this change benchmarking gives a 6% speed
improvement that can be consistently measured.
2014-05-07 16:17:11 +02:00
antirez 3a1a0c3240 Scripting: luaRedisGenericCommand() fast path for buffer-only replies.
When the reply is only contained in the client static output buffer, use
a fast path avoiding the dynamic allocation of an SDS string to
concatenate the client reply objects.
2014-05-07 16:17:11 +02:00
antirez d70281575c Define HAVE_ATOMIC for clang. 2014-05-07 16:17:11 +02:00
antirez 67e350367d Scripting: simpler reply buffer creation in luaRedisGenericCommand().
It if faster to just create the string with a single sdsnewlen() call.
If c->bufpos is zero, the call will simply be like sdsemtpy().
2014-05-07 16:17:11 +02:00
Matt Stancliff b06de30dd7 Add test for deleting an expired key
Verify proper expire-before-delete behavior.

This test passes with the expire-before-delete commit and fails
without it.
2014-04-23 16:14:40 +02:00
Matt Stancliff e9d0d86dbf Check key expiration before deleting
Deleting an expired key should return 0, not success.

Fixes #1648
2014-04-23 16:14:40 +02:00
Glauber Costa 96b5dc0b98 fix null pointer access with no file pointer
I happen to be working on a system that lacks urandom. While the code does try
to handle this case and artificially create some bytes if the file pointer is
empty, it does try to close it unconditionally, leading to a segfault.
2014-04-23 15:35:12 +02:00
Salvatore Sanfilippo 3d2af57974 Merge pull request #1700 from nirvdrum/patch-1
Fixed typos.
2014-04-23 11:55:50 +02:00
antirez 35252037fd Missing return REDIS_ERR added to processMultibulkBuffer().
When we set a protocol error we should return with REDIS_ERR to let the
caller know it should stop processing the client.

Bug found in a code auditing related to issue #1699.
2014-04-23 10:21:26 +02:00
Kevin Menard 662f505727 Fixed typos. 2014-04-22 10:51:19 -04:00
antirez de7ebf83e1 redis-cli help.h updated. 2014-04-22 16:15:01 +02:00
antirez 91dcc5b12f generate-command-help.rb updated with new hyperloglog group. 2014-04-22 16:15:01 +02:00
antirez 26375709c5 Redis 2.8.9. 2014-04-22 10:14:57 +02:00
antirez c5ee48a9d9 Fuzzy test for ZREMRANGEBYLEX added. 2014-04-18 16:16:20 +02:00
antirez be145453c6 ZREMRANGEBYLEX memory leak removed calling zslFreeLexRange(). 2014-04-18 16:16:20 +02:00
antirez 115782b1d7 PFCOUNT multi-key test added. 2014-04-18 16:16:20 +02:00
antirez 2cd20d3c4c Speedup hllRawSum() processing 8 bytes per iteration.
The internal HLL raw encoding used by PFCOUNT when merging multiple keys
is aligned to 8 bits (1 byte per register) so we can exploit this to
improve performances by processing multiple bytes per iteration.

In benchmarks the new code was several times faster with HLLs with many
registers set to zero, while no slowdown was observed with populated
HLLs.
2014-04-18 16:16:20 +02:00
antirez 3f5cd0c660 Speedup SUM(2^-reg[m]) in HyperLogLog computation.
When the register is set to zero, we need to add 2^-0 to E, which is 1,
but it is faster to just add 'ez' at the end, which is the number of
registers set to zero, a value we need to compute anyway.
2014-04-18 16:16:20 +02:00
antirez f0eca148ff PFCOUNT support for multi-key union. 2014-04-18 16:16:20 +02:00
antirez d25fd8be64 HyperLogLog low level merge extracted from PFMERGE. 2014-04-18 16:16:20 +02:00
antirez 4a26648a22 ZREMRANGEBYLEX implemented. 2014-04-18 16:16:20 +02:00
antirez 3fc22561e7 Always pass sorted set range objects by reference. 2014-04-18 16:16:04 +02:00
antirez ea0af9ace3 ZREMRANGE* commands refactored into a single generic function. 2014-04-18 16:14:38 +02:00
antirez 019e6f1889 Sorted set lex ranges stress tester. 2014-04-18 16:14:38 +02:00
antirez 9c9aef24a2 Basic ZRANGEBYLEX / ZLEXCOUNT tests. 2014-04-17 00:08:38 +02:00
antirez 11f89497b4 Pass by pointer and release of lex ranges.
Given that the code was written with a 2 years pause... something
strange happened in the middle. So there was no function to free a
lex range min/max objects, and in some places the range was passed by
value.
2014-04-17 00:08:37 +02:00
antirez d975bb587f ZLEXCOUNT implemented.
Like ZCOUNT for lexicographical ranges.
2014-04-16 15:26:28 +02:00
antirez 71dfb87f76 More HyperLogLog tests. 2014-04-16 15:26:28 +02:00
antirez f2e59eae44 HyperLogLog invalid representation error code set to INVALIDOBJ. 2014-04-16 15:26:28 +02:00
antirez 469b7c5168 PFDEBUG TODENSE added.
Converts HyperLogLogs from sparse to dense. Used for testing.
2014-04-16 15:26:28 +02:00
antirez 614fcd491e User-defined switch point between sparse-dense HLL encodings. 2014-04-16 15:26:28 +02:00
antirez 08da2b79fc PFSELFTEST improved with sparse encoding checks. 2014-04-16 15:26:27 +02:00
antirez 5a0bab2a8f PFDEBUG ENCODING added. 2014-04-16 15:26:27 +02:00
antirez 05563310f4 Set HLL_SPARSE_MAX to 3000.
After running a few benchmarks, 3000 looks like a reasonable value to
keep HLLs with a few thousand elements small while the CPU cost is
still not huge.

This covers all the cases where the dense representation would use N
orders of magnitude more space, like in the case of many HLLs with
carinality of a few tens or hundreds.

It is not impossible that in the future this gets user configurable,
however it is easy to pick an unreasoable value just looking at savings
in the space dimension without checking what happens in the time
dimension.
2014-04-16 15:26:27 +02:00
antirez e77a3e40f2 Error message for invalid HLL objects unified. 2014-04-16 15:26:27 +02:00
antirez c4b2a7a7f4 PFMERGE fixed to work with sparse encoding. 2014-04-16 15:26:27 +02:00
antirez ee682c4c48 Mark PFDEBUG as write command in the commands table.
It is safer since it is able to have side effects.
2014-04-16 15:26:27 +02:00
antirez 9ef8ef410b Correctly replicate PFDEBUG GETREG.
Even if it is a debugging command, make sure that when it forces a
change in encoding, the command is propagated.
2014-04-16 15:26:27 +02:00
antirez 91e05e1618 Added assertion in hllSparseAdd() when promotion to dense occurs.
If we converted to dense, a register must be updated in the dense
representation.
2014-04-16 15:26:27 +02:00
antirez 68a4bf70e0 hllSparseAdd(): speed optimization.
Mostly by reordering opcodes check conditional by frequency of opcodes
in larger sparse-encoded HLLs.
2014-04-16 15:26:27 +02:00
antirez 21e328a6d7 Detect corrupted sparse HLLs in hllSparseSum(). 2014-04-16 15:26:27 +02:00
antirez 4244aa29f9 hllSparseAdd(): faster code removing conditional.
Bottleneck found profiling. Big run time improvement found when testing
after the change.
2014-04-16 15:26:27 +02:00
antirez 4cd11077d2 Comment typo in hllSparseAdd(). first -> fits. 2014-04-16 15:26:27 +02:00
antirez 08183cfef5 Merge adjacent VAL opcodes in hllSparseAdd().
As more values are added splitting ZERO or XZERO opcodes, try to merge
adjacent VAL opcodes if they have the same value.
2014-04-16 15:26:27 +02:00
antirez d428569d13 More robust HLL_SPARSE macros protecting 'p' with parens.
Now the macros will work with arguments such as "ptr+1".
2014-04-16 15:26:27 +02:00
antirez 2d1e35a58a hllSparseAdd() opcode seek stop condition fixed. 2014-04-16 15:26:27 +02:00
antirez 29c7d1974b Fixed error message generation in PFDEBUG GETREG.
Bulk length for registers was emitted too early, so if there was a bug
the reply looked like a long array with just one element, blocking the
client as result.
2014-04-16 15:26:27 +02:00
antirez 023f2f7a10 Fixed memmove() count in hllSparseAdd(). 2014-04-16 15:26:27 +02:00
antirez 91391760ca hllSparseAdd(): more correct dense conversion conditional.
We want to promote if the total string size exceeds the resulting size
after the upgrade.
2014-04-16 15:26:27 +02:00
antirez 5d87160f8a hllSparseToDense(): sanity check added.
The function checks if all the HLL_REGISTERS were processed during the
convertion from sparse to dense encoding, returning REDIS_OK or
REDIS_ERR to signal a corruption problem.

A bug in PFDEBUG GETREG was fixed: when the object is converted to the
dense representation we need to reassign the new pointer to the header
structure pointer.
2014-04-16 15:26:27 +02:00
antirez a18e7afbe8 PFDEBUG DECODE added.
Provides a human readable description of the opcodes composing a
run-length encoded HLL (sparse encoding).
The command is only useful for debugging / development tasks.
2014-04-16 15:26:27 +02:00
antirez 5c86d9e7b0 PFDEBUG added, PFGETREG removed.
PFDEBUG will be the interface to do debugging tasks with a key
containing an HLL object.
2014-04-16 15:26:27 +02:00
antirez 0e0ddcc4fa hllSparseToDense API changed to take ref to object.
The new API takes directly the object doing everything needed to
turn it into a dense representation, including setting the new
representation as object->ptr.
2014-04-16 15:26:27 +02:00
antirez 533235f7e3 hllSparseAdd() sanity check for span != 0 added. 2014-04-16 15:26:27 +02:00
antirez f12fb49bb2 Fix hllSparseAdd() new sequence replacement when next is NULL.
sdsIncrLen() must be called anyway even if we are replacing the last
oppcode of the sparse representation.
2014-04-16 15:26:27 +02:00
antirez a3e2adc0ae Fix seqlen computation in hllSparseAdd(). 2014-04-16 15:26:27 +02:00
antirez eaa73ef73a Abstract hllSparseAdd() / hllDenseAdd() via hllAdd(). 2014-04-16 15:26:27 +02:00
antirez 95de3eaea0 hllSparseSum(): multiply 1 * runlen for zero entries. 2014-04-16 15:26:27 +02:00
antirez c277169ed7 Macro HLL_SPARSE_XZERO_LEN fixed. 2014-04-16 15:26:27 +02:00
antirez b6ca71a40a Fix HLL sparse object creation #2.
Two vars initialized to wrong values in createHLLObject().
2014-04-16 15:26:27 +02:00
antirez 8b39fee4c8 Increment pointer while iterating sparse HLL object. 2014-04-16 15:26:27 +02:00
antirez c1ddead779 Fix HLL sparse object creation.
The function didn't considered the fact that each XZERO opcode is
two bytes.
2014-04-16 15:26:27 +02:00
antirez 0a7ba6faef Create HyperLogLog objects with sparse encoding. 2014-04-16 15:26:27 +02:00
antirez 8e889728bf HyperLogLog sparse to dense conversion function. 2014-04-16 15:26:27 +02:00
antirez aadd513160 HyperLogLog sparse representation initial implementation.
Code never tested, but the basic layout is shaped in this commit.
Also missing:

1) Sparse -> Dense conversion function.
2) New HLL object creation using the sparse representation.
3) Implementation of PFMERGE for the sparse representation.
2014-04-16 15:26:27 +02:00
antirez e5e5facba1 hllCount() refactored to support multiple representations. 2014-04-16 15:26:27 +02:00
antirez 88847a216f hllAdd() refactored into two functions.
Also dense representation access macro renamed accordingly.
2014-04-16 15:26:27 +02:00
antirez 21c97e71cd HyperLogLog refactoring to support different encodings.
Metadata are now placed at the start of the representation as an header.
There is a proper structure to access the representation.
Still work to do in order to truly abstract the implementation from the
representation, commands still work assuming dense representation.
2014-04-16 15:26:27 +02:00
antirez ed96e115d0 HyperLogLog sparse representation slightly modified.
After running a few simulations with different alternative encodings,
it was found that the VAL opcode performs better using 5 bits for the
value and 2 bits for the run length, at least for cardinalities in the
range of interest.
2014-04-16 15:26:27 +02:00
antirez 93c75f1bf4 HyperLogLog sparse representation description and macros. 2014-04-16 15:26:26 +02:00
antirez 2b1385b9ec Add casting to match printf format.
adjustOpenFilesLimit() and clusterUpdateSlotsWithConfig() that were
assuming uint64_t is the same as unsigned long long, which is true
probably for all the systems out there that we target, but still GCC
emitted a warning since technically they are two different types.
2014-04-16 15:26:22 +02:00
antirez 12d1d18b59 ZRANGEBYLEX and ZREVRANGEBYLEX implementation. 2014-04-16 15:26:09 +02:00
antirez a10bfded15 PFCOUNT: always unshare/decode the object.
This will be a non-op most of the times since the object will be
unshared / decoded, however it is more technically correct to start this
way since the object may be decoded even in the read-only code path.
2014-04-16 15:26:09 +02:00
antirez ee764b0f8d Changed HyperLogLog hash seed to a non-zero value.
Using a seed of zero has the side effect of having the empty string
hashing to what is a very special case in the context of HyperLogLog: a
very long run of zeroes.

This did not influenced the correctness of the result with 16k registers
because of the harmonic mean, but still it is inconvenient that a so
obvious value maps to a so special hash.

The seed 0xadc83b19 is used instead, which is the first 64 bits of the
SHA1 of the empty string.

Reference: issue #1657.
2014-04-16 15:26:09 +02:00
antirez 091f5677a0 Initial HyperLogLog tests. 2014-04-16 15:26:09 +02:00
antirez a97675923d Return "WRONGTYPE" error on PF* type mismatch. 2014-04-16 15:26:09 +02:00
antirez faa7f259cc Fix PFADD infinite loop.
We need to guarantee that the last bit is 1, otherwise an element may
hash to just zeroes with probability 1/(2^64) and trigger an infinite
loop.

See issue #1657.
2014-04-16 15:26:09 +02:00
antirez 8ae6887091 Make hll-gnuplot-graph.rb callable from cli. 2014-04-16 15:26:09 +02:00
antirez bdd8701c60 Remove HyperLogLog type checking duplicated code. 2014-04-16 15:26:09 +02:00
antirez b3baa51403 PFGETREG added for testing purposes.
The new command allows to get a dump of the registers stored
into an HyperLogLog data structure for testing / debugging purposes.
2014-04-16 15:26:09 +02:00
antirez 348d5ea246 PFCOUNT: unshare the object when cached cardinality is modified. 2014-04-16 15:26:09 +02:00
antirez ba0dc0be2c PFSELFTEST improved to test the approximation error. 2014-04-16 15:26:09 +02:00
antirez c2cd48718d hll-gnuplot-graph.rb improved with new filter.
The function to generate graphs is also more flexible as now includes
step and max value. The step of the samples generation function is no
longer limited to min step of 1000.
2014-04-16 15:26:09 +02:00
antirez ff551fad24 HyperLogLog: added magic / version.
This will allow future changes like compressed representations.
Currently the magic is not checked for performance reasons but this may
change in the future, for example if we add new types encoded in strings
that may have the same size of HyperLogLogs.
2014-04-16 15:26:09 +02:00
Raymond Myers e08100a9ba Fixed pfadd/pfcount commands emitting hll* events instead of pf* events 2014-04-16 15:26:09 +02:00
Raymond Myers 23acae6e1c Change HLL* to PF* in error messages 2014-04-16 15:26:09 +02:00
antirez e3773b0a0e Include redis.h before other stuff in hyperloglog.c.
Otherwise fmacros.h is included later and this may break compilation on
different systems.
2014-04-16 15:26:09 +02:00
antirez badf23f57b HyperLogLog API prefix modified from "P" to "PF".
Using both the initials of Philippe Flajolet instead of just "P".
2014-04-16 15:26:03 +02:00
antirez 6e71459c73 Makefile.dep updated with hyperloglog.o deps. 2014-04-16 15:25:28 +02:00
antirez 2ee2bec2d6 HyperLogLog: make API use the P prefix in honor of Philippe Flajolet. 2014-04-16 15:24:15 +02:00
antirez 1e90c0066c HLLMERGE fixed by adding a... missing loop! 2014-04-16 15:23:21 +02:00
antirez be7eef911b hll-gnuplot-graph.rb: added new filter "all". 2014-04-16 15:23:21 +02:00
antirez f2e707b679 HyperLogLog apply bias correction using a polynomial.
Better results can be achieved by compensating for the bias of the raw
approximation just after 2.5m (when LINEARCOUNTING is no longer used) by
using a polynomial that approximates the bias at a given cardinality.

The curve used was found using this web page:

    http://www.xuru.org/rt/PR.asp

That performs polynomial regression given a set of values.
2014-04-16 15:23:21 +02:00
antirez be7fe2b92b HLLMERGE implemented.
Merge N HLL data structures by selecting the max value for every
M[i] register among the set of HLLs.
2014-04-16 15:23:21 +02:00
antirez 39dd1f648d HLLCOUNT is technically a write command
When we update the cached value, we need to propagate the command and
signal the key as modified for WATCH.
2014-04-16 15:23:21 +02:00
antirez f07b1514bb HLLADD: propagate write when only variable name is given.
The following form is given:

    HLLADD myhll

No element is provided in the above case so if 'myhll' var does not
exist the result is to just create an empty HLL structure, and no update
will be performed on the registers.

In this case, the DB should still be set dirty and the command
propagated.
2014-04-16 15:23:21 +02:00
antirez 32e9996846 hll-gnuplot-graph.rb: Use |error| when filter is :max 2014-04-16 15:23:21 +02:00
antirez 9faa80ee8c Ignore txt files inside utils/hyperloglog.
Those are generated to trace graphs using gnuplot.
2014-04-16 15:23:21 +02:00
antirez aad5959ddc HyperLogLog: use LINEARCOUNTING up to 3m.
The HyperLogLog original paper suggests using LINEARCOUNTING for
cardinalities < 2.5m, however for P=14 the median / max error
curves show that a value of '3' is the best pick for m = 16384.
2014-04-16 15:23:21 +02:00
antirez d6ce659ffd hll-gnuplot-graph.rb added to plot HyperLogLog error graphs. 2014-04-16 15:23:21 +02:00
antirez c3df5965f3 HyperLogLog approximated cardinality caching.
The more we add elements to an HyperLogLog counter, the smaller is
the probability that we actually update some register.

From this observation it is easy to see how it is possible to use
caching of a previously computed cardinality and reuse it to serve
HLLCOUNT queries as long as no register was updated in the data
structure.

This commit does exactly this by using just additional 8 bytes for the
data structure to store a 64 bit unsigned integer value cached
cardinality. When the most significant bit of the 64 bit integer is set,
it means that the value computed is no longer usable since at least a
single register was modified and we need to recompute it at the next
call of HLLCOUNT.

The value is always stored in little endian format regardless of the
actual CPU endianess.
2014-04-16 15:23:21 +02:00
antirez 9e178afae9 String value unsharing refactored into proper function.
All the Redis functions that need to modify the string value of a key in
a destructive way (APPEND, SETBIT, SETRANGE, ...) require to make the
object unshared (if refcount > 1) and encoded in raw format (if encoding
is not already REDIS_ENCODING_RAW).

This was cut & pasted many times in multiple places of the code. This
commit puts the small logic needed into a function called
dbUnshareStringValue().
2014-04-16 15:22:56 +02:00
antirez 1c795db9d0 Use endian neutral hash function for HyperLogLog.
We need to be sure that you can save a dataset in a Redis instance,
reload it in a different architecture, and continue to count in the same
HyperLogLog structure.

So 32 and 64 bit, little or bit endian, must all guarantee to output the
same hash for the same element.
2014-04-16 15:19:40 +02:00
antirez 7f30998432 HyperLogLog internal representation modified.
The new representation is more obvious, starting from the LSB of the
first byte and using bits going to MSB, and passing to next byte as
needed.

There was also a subtle error: first two bits were unused, everything
was carried over on the right of two bits, even if it worked because of
the code requirement of always having a byte more at the end.

During the rewrite the code was made safer trying to avoid undefined
behavior due to shifting an uint8_t for more than 8 bits.
2014-04-16 15:19:40 +02:00
antirez 430d0ade75 Remove a few useless operations from hllCount() fast path. 2014-04-16 15:19:40 +02:00
antirez 83650a72a1 HLLCOUNT 3x faster taking fast path for default params. 2014-04-16 15:19:40 +02:00
antirez 3afd25ecc7 Use processor base types in HLL_(GET|SET)_REGISTER.
This speedups the macros by a noticeable factor.
2014-04-16 15:19:40 +02:00
antirez 0ffd5e4a30 HyperLogLog: use precomputed table for 2^(-M[i]). 2014-04-16 15:19:40 +02:00
antirez ca6973d3b6 hll-err.rb: speedup using pipelining. 2014-04-16 15:19:40 +02:00
antirez 2500f2a334 hll-err.rb added to test error rate of Redis HyperLogLog. 2014-04-16 15:19:40 +02:00
antirez 42592a7faa HyperLogLog algorithm fixed in two ways.
There was an error in the computation of 2^register, and the sequence of
zeroes computed after the hashing did not included the "1".
2014-04-16 15:19:40 +02:00
antirez 3035075b2c HLLCOUNT implemented. 2014-04-16 15:19:40 +02:00
antirez 959d0f012a HLLADD implemented. 2014-04-16 15:19:33 +02:00
antirez ac29bb2ae4 hllAdd() low level HyperLogLog "add" implemented. 2014-04-16 15:19:05 +02:00
antirez 2a91f548c4 HyperLogLog: redefine constants using "P". 2014-04-16 15:19:05 +02:00
antirez 3f159bbd44 HLL_SET_REGISTER fixed.
There was an error in the first version of the macro.
Now the HLLSELFTEST test reports success.
2014-04-16 15:19:05 +02:00
antirez a18e880bf1 Use REDIS_HLL_REGISTER_MAX when possible. 2014-04-16 15:19:05 +02:00
antirez 00409d8742 HLL_(SET|GET)_REGISTER types fixed. 2014-04-16 15:19:05 +02:00
antirez de3f821af6 HLLSELFTEST command implemented.
To test the bitfield array of counters set/get macros from the Redis Tcl
suite is hard, so a specialized command that is able to test the
internals was developed.
2014-04-16 15:18:57 +02:00
antirez 02d88fb201 HyperLogLog: initial sketch of registers access. 2014-04-16 15:17:26 +02:00
antirez ddbdd3b0df Document TTL changes in the 2.6 -> 2.8 migration doc. 2014-03-27 15:53:14 +01:00
antirez d2e59c2715 Redis 2.8.8. 2014-03-25 11:30:42 +01:00
antirez 99fc582f8c Test: do not complain when "leaks" can't run because process died. 2014-03-25 09:34:11 +01:00
antirez 3580bb485a adjustOpenFilesLimit() refactoring.
In this commit:
* Decrement steps are semantically differentiated from the reserved FDs.
  Previously both values were 32 but the meaning was different.
* Make it clear that we save setrlimit errno.
* Don't explicitly handle wrapping of 'f', but prevent it from
  happening.
* Add comments to make the function flow more readable.

This integrates PR #1630
2014-03-25 09:07:21 +01:00
Matt Stancliff c3510af1c0 Fix potentially incorrect errno usage
errno may be reset by the previous call to redisLog, so capture
the original value for proper error reporting.
2014-03-25 09:07:21 +01:00
Matt Stancliff 771f8ad0e7 Add REDIS_MIN_RESERVED_FDS define for open fds
Also update the original REDIS_EVENTLOOP_FDSET_INCR to
include REDIS_MIN_RESERVED_FDS. REDIS_EVENTLOOP_FDSET_INCR
exists to make sure more than (maxclients+RESERVED) entries
are allocated, but we can only guarantee that if we include
the current value of REDIS_MIN_RESERVED_FDS as a minimum
for the INCR size.
2014-03-25 09:07:21 +01:00
Matt Stancliff 3ce742d1d5 Fix infinite loop on startup if ulimit too low
Fun fact: rlim_t is an unsigned long long on all platforms.

Continually subtracting from a rlim_t makes it get smaller
and smaller until it wraps, then you're up to 2^64-1.

This was causing an infinite loop on Redis startup if
your ulimit was extremely (almost comically) low.

The case of (f > oldlimit) would never be met in a case like:

    f = 150
    while (f > 20) f -= 128

Since f is unsigned, it can't go negative and would
take on values of:

    Iteration 1: 150 - 128 => 22
    Iteration 2:  22 - 128 => 18446744073709551510
    Iterations 3-∞: ...

To catch the wraparound, we use the previous value of f
stored in limit.rlimit_cur.  If we subtract from f and
get a larger number than the value it had previously,
we print an error and exit since we don't have enough
file descriptors to help the user at this point.

Thanks to @bs3g for the inspiration to fix this problem.
Patches existed from @bs3g at antirez#1227, but I needed to repair a few other
parts of Redis simultaneously, so I didn't get a chance to use them.
2014-03-25 09:07:21 +01:00
Matt Stancliff b59585396b Improve error handling around setting ulimits
The log messages about open file limits have always
been slightly opaque and confusing.  Here's an attempt to
fix their wording, detail, and meaning.  Users will have a
better understanding of how to fix very common problems
with these reworded messages.

Also, we handle a new error case when maxclients becomes less
than one, essentially rendering the server unusable.  We
now exit on startup instead of leaving the user with a server
unable to handle any connections.

This fixes antirez#356 as well.
2014-03-25 09:07:21 +01:00
Matt Stancliff 6826af1b50 Replace magic 32 with REDIS_EVENTLOOP_FDSET_INCR
32 was the additional number of file descriptors Redis
would reserve when managing a too-low ulimit.  The
number 32 was in too many places statically, so now
we use a macro instead that looks more appropriate.

When Redis sets up the server event loop, it uses:
    server.maxclients+REDIS_EVENTLOOP_FDSET_INCR

So, when reserving file descriptors, it makes sense to
reserve at least REDIS_EVENTLOOP_FDSET_INCR FDs instead
of only 32.  Currently, REDIS_EVENTLOOP_FDSET_INCR is
set to 128 in redis.h.

Also, I replaced the static 128 in the while f < old loop
with REDIS_EVENTLOOP_FDSET_INCR as well, which results
in no change since it was already 128.

Impact: Users now need at least maxclients+128 as
their open file limit instead of maxclients+32 to obtain
actual "maxclients" number of clients.  Redis will carve
the extra REDIS_EVENTLOOP_FDSET_INCR file descriptors it
needs out of the "maxclients" range instead of failing
to start (unless the local ulimit -n is too low to accomidate
the request).
2014-03-25 09:07:21 +01:00
Matt Stancliff 611372fa97 Fix maxclients error handling
Everywhere in the Redis code base, maxclients is treated
as an int with (int)maxclients or `maxclients = atoi(source)`,
so let's make maxclients an int.

This fixes a bug where someone could specify a negative maxclients
on startup and it would work (as well as set maxclients very high)
because:

    unsigned int maxclients;
    char *update = "-300";
    maxclients = atoi(update);
    if (maxclients < 1) goto fail;

But, (maxclients < 1) can only catch the case when maxclients
is exactly 0.  maxclients happily sets itself to -300, which isn't
-300, but rather 4294966996, which isn't < 1, so... everything
"worked."

maxclients config parsing checks for the case of < 1, but maxclients
CONFIG SET parsing was checking for case of < 0 (allowing
maxclients to be set to 0).  CONFIG SET parsing is now updated to
match config parsing of < 1.

It's tempting to add a MINIMUM_CLIENTS define, but... I didn't.

These changes were inspired by antirez#356, but this doesn't
fix that issue.
2014-03-25 09:07:21 +01:00
Matt Stancliff 3bd3240660 Sentinel: remove variable causing warning
GCC-4.9 warned about this, but clang didn't.

This commit fixes warning:
sentinel.c: In function 'sentinelReceiveHelloMessages':
sentinel.c:2156:43: warning: variable 'master' set but not used [-Wunused-but-set-variable]
     sentinelRedisInstance *ri = c->data, *master;
2014-03-25 08:06:04 +01:00
antirez 79349affd5 Fixed undefined variable value with certain code paths.
In sentinelFlushConfig() fd could be undefined when the following if
statement was true:

        if (rewrite_status == -1) goto werr;

This could cause random file descriptors to get closed.
2014-03-25 08:03:11 +01:00
Matt Stancliff 80dec5e4df Sentinel: Notify user when config can't be saved 2014-03-25 08:03:07 +01:00
Jan-Erik Rediger a2ec9a9068 Small typo fixed 2014-03-25 08:02:56 +01:00
Matt Stancliff 88c6c669c3 Fix data loss when save AOF/RDB with no free space
Previously, the (!fp) would only catch lack of free space
under OS X.  Linux waits to discover it can't write until
it actually writes contents to disk.

(fwrite() returns success even if the underlying file
has no free space to write into.  All the errors
only show up at flush/sync/close time.)

Fixes antirez/redis#1604
2014-03-24 21:56:22 +01:00
Jan-Erik Rediger 6182dad37c Finally fix the install_server.sh script.
Includes changes from a dozen bug reports and pull requests.
Was tested on Ubuntu, Debian and CentOS.
2014-03-24 18:12:58 +01:00
antirez 4ebc7e3738 Sample and cache RSS in serverCron().
Obtaining the RSS (Resident Set Size) info is slow in Linux and OSX.
This slowed down the generation of the INFO 'memory' section.

Since the RSS does not require to be a real-time measurement, we
now sample it with server.hz frequency (10 times per second by default)
and use this value both to show the INFO rss field and to compute the
fragmentation ratio.

Practically this does not make any difference for memory profiling of
Redis but speeds up the INFO call significantly.
2014-03-24 12:03:44 +01:00
antirez 72257f4b87 sdscatvprintf(): Try to use a static buffer.
For small content the function now tries to use a static buffer to avoid
a malloc/free cycle that is too costly when the function is used in the
context of performance critical code path such as INFO output generation.

This change was verified to have positive effects in the execution speed
of the INFO command.
2014-03-24 10:23:58 +01:00
antirez 7054f2a029 Cache uname() output across INFO calls.
Uname was profiled to be a slow syscall. It produces always the same
output in the context of a single execution of Redis, so calling it at
every INFO output generation does not make too much sense.

The uname utsname structure was modified as a static variable. At the
same time a static integer was added to check if we need to call uname
the first time.
2014-03-24 10:02:58 +01:00
antirez 7a20f09647 sdscatvprintf(): guess buflen using format length.
sdscatvprintf() uses a loop where it tries to output the formatted
string in a buffer of the initial length, if there was not enough room,
a buffer of doubled size is tried and so forth.

The initial guess for the buffer length was very poor, an hardcoded
"16". This caused the printf to be processed multiple times without a
good reason. Given that printf functions are already not fast, the
overhead was significant.

The new heuristic is to use a buffer 4 times the length of the format
buffer, and 32 as minimal size. This appears to be a good balance for
typical uses of the function inside the Redis code base.

This change improved INFO command performances 3 times.
2014-03-24 09:45:18 +01:00
antirez 001775f5fb Use 24 bits for the lru object field and improve resolution.
There were 2 spare bits inside the Redis object structure that are now
used in order to enlarge 4x the range of the LRU field.

At the same time the resolution was improved from 10 to 1 second: this
still provides 194 days before the LRU counter overflows (restarting from
zero).

This is not a problem since it only causes lack of eviction precision for
objects not touched for a very long time, and the lack of precision is
only temporary.
2014-03-21 11:19:28 +01:00
antirez c68189a19f Specify lruclock in redisServer structure via REDIS_LRU_BITS.
The padding field was totally useless: removed.
2014-03-21 11:17:52 +01:00
antirez ff8c81873a Set LRU parameters via REDIS_LRU_BITS define. 2014-03-21 11:17:36 +01:00
antirez e3b71a1c02 Unify stats reset for CONFIG RESETSTAT / initServer().
Now CONFIG RESETSTAT makes sure to reset all the fields, and in the
future it will be simpler to avoid missing new fields.
2014-03-21 11:16:12 +01:00
antirez 0937377aa2 Sentinel: sentinelRefreshInstanceInfo() minor refactoring.
Test sentinel.tilt condition on top and return if it is true.
This allows to remove the check for the tilt condition in the remaining
code paths of the function.
2014-03-21 11:16:12 +01:00
antirez 686839b477 Sentinel test: 02 unit better coverage + refactoring. 2014-03-21 11:16:12 +01:00
antirez 6d0e408a27 Sentinel test: foreach_instance_id implements 'break'. 2014-03-21 11:16:12 +01:00
antirez ba2edc4191 Sentinel: instance_is_killed proc added to sentinel.tcl. 2014-03-21 11:16:12 +01:00
antirez 9c2063fb8f Sentinel: propagate down-after-ms changes to slaves and sentinels. 2014-03-21 11:16:12 +01:00
antirez ffa8f479a5 Sentinel: down-after-milliseconds is not master-specific.
addReplySentinelRedisInstance() modified so that this field is displayed
for all the kind of instances: Sentinels, Masters, Slaves.
2014-03-21 11:16:12 +01:00
antirez 42091a79bb Sentinel failure detection implementation improved.
Failure detection in Sentinel is ping-pong based. It used to work by
remembering the last time a valid PONG reply was received, and checking
if the reception time was too old compared to the current current time.

PINGs were sent at a fixed interval of 1 second.

This works in a decent way, but does not scale well when we want to set
very small values of "down-after-milliseconds" (this is the node
timeout basically).

This commit reiplements the failure detection making a number of
changes. Some changes are inspired to Redis Cluster failure detection
code:

* A new last_ping_time field is added in representation of instances.
  If non zero, we have an active ping that was sent at the specified
  time. When a valid reply to ping is received, the field is zeroed
  again.
* last_ping_time is not reset when we reconnect the link or send a new
  ping, so from our point of view it represents the time we started
  waiting for the instance to reply to our pings without receiving a
  reply.
* last_ping_time is now used in order to check if the instance is
  timed out. This means that we can have a node timeout of 100
  milliseconds and yet the system will work well since the new check is
  not bound to the period used to send pings.
* Pings are now sent every second, or often if the value of
  down-after-milliseconds is less than one second. With a lower limit of
  10 HZ ping frequency.
* Link reconnection code was improved. This is used in order to try to
  reconnect the link when we are at 50% of the node timeout without a
  valid reply received yet. However the old code triggered unnecessary
  reconnections when the node timeout was very small. Now that should be
  ok.

The new code passes the tests but more testing is needed and more unit
tests stressing the failure detector, so currently this is merged only
in the unstable branch.
2014-03-21 11:16:11 +01:00
antirez 38241c4b1e Sentinel: use CLIENT SETNAME when connecting to Redis.
This makes debugging / monitoring of Sentinels simpler since you can
identify sentinels in CLIENT LIST output of Redis instances.
2014-03-21 11:16:11 +01:00
Matt Stancliff 9de0755869 Fix segfault from accessing array out of bounds
argc == 2; argv[2] == crash
2014-03-14 22:57:10 +01:00
antirez a31a0b4336 Sentinel: be safe under crash-recovery assumptions.
Sentinel's main safety argument is that there are no two configurations
for the same master with the same version (configuration epoch).

For this to be true Sentinels require to be authorized by a majority.
Additionally Sentinels require to do two important things:

* Never vote again for the same epoch.
* Never exchange an old vote for a fresh one.

The first prerequisite, in a crash-recovery system model, requires to
persist the master->leader_epoch on durable storage before to reply to
messages. This was not the case.

We also make sure to persist the current epoch in order to never reply
to stale votes requests from other Sentinels, after a recovery.

The configuration is persisted by making use of fsync(), this is
considered in the context of this code a good enough guarantee that
after a restart our durable state is restored, however this may not
always be the case depending on the kind of hardware and operating
system used.
2014-03-14 22:57:06 +01:00
antirez 6b0e36ffc6 Sentinel: fake PUBLISH command to receive HELLO messages.
Now the way HELLO messages are received is unified.
Now it is no longer needed for Sentinels to converge to the higher
configuration for a master to be able to chat via some Redis instance,
the are able to directly exchanges configurations.

Note that this commit does not include the (trivial) change needed to
send HELLO messages to Sentinel instances as well, since for an error I
committed the change in the previous commit that refactored hello
messages processing into a separated function.
2014-03-14 11:04:54 +01:00
antirez bd48ff69b0 Sentinel: HELLO processing refactored into sentinelProcessHelloMessage(). 2014-03-14 10:56:56 +01:00
antirez 3703112671 Linenoise updated, multiline mode enabled in redis-cli. 2014-03-13 15:12:04 +01:00
zhanghailei ccd2c18c1c According to context,the size should be 16 rather than 64 2014-03-11 10:12:19 +01:00
zhanghailei daa5d2a6c9 FIXED a typo more thank should be more than 2014-03-11 10:12:13 +01:00
zhanghailei ca0720b694 refer to updateLRUClock's comment REDIS_LRU_CLOCK_MAX is 22 bits,but #define REDIS_LRU_CLOCK_MAX ((1<<21)-1) only 21 bits 2014-03-11 10:12:00 +01:00
Matt Stancliff ffe742f92a Reset op_sec_last_sample_ops when reset requested
This value needs to be set to zero (in addition to
stat_numcommands) or else people may see
a negative operations per second count after they
run CONFIG RESETSTAT.

Fixes antirez/redis#1577
2014-03-11 10:10:54 +01:00
antirez 6fb9e2d717 Typo in sentinel.conf, exists -> exits. 2014-03-11 10:10:30 +01:00
antirez eb9e1526e6 DEBUG ERROR implemented.
The new "error" subcommand of the DEBUG command can reply with an user
selected error, specified as its sole argument:

    DEBUG ERROR "LOADING please wait..."

The error is generated just prefixing the command argument with a "-"
character, and replacing newlines with spaces (since error replies can't
include newlines).

The goal of the command is to help in Client libraries unit tests by
making simple to simulate a command call triggering a given error.
2014-03-10 23:04:37 +01:00
Matt Stancliff a6970570e3 Fix return value check for anetTcpAccept
anetTcpAccept returns ANET_ERR, not AE_ERR.

This isn't a physical error since both ANET_ERR
and AE_ERR are -1, but better to be consistent.
2014-03-10 15:47:20 +01:00
antirez fe0ab7d234 Fixed memory leak in SORT LIMIT option argument parsing on error. 2014-03-10 15:45:29 +01:00
49 changed files with 4646 additions and 734 deletions
+76
View File
@@ -14,6 +14,79 @@ HIGH: There is a critical bug that may affect a subset of users. Upgrade!
CRITICAL: There is a critical bug affecting MOST USERS. Upgrade ASAP.
--------------------------------------------------------------------------------
--[ Redis 2.8.10 ] Release date: 5 Jun 2014
# UPGRADE URGENCY: HIGH if you use min-slaves-to-write option.
* [FIX] IMPORTANT! A min-slaves-to-write option active in a slave totally
prevented the slave from acception the master stream of commands.
This release includes testes for min-slaves-to-write, and a fix
for this issue.
* [FIX] Sometimes DEL returned 1 for already expired keys. Fixed.
* [FIX] Fix test false positive because new osx 'leaks' output.
* [FIX] PFCOUNT HLL cache invalidation fixed: no wrong value was reported
but the cache was not used at all, leading to lower performances.
* [FIX] Accept(2) multiple clients per readable-event invocation, and better
processing of I/O while loading or busy running a timedout script.
Basically now the LOADING / BUSY errors are reported at a decent
speed.
* [FIX] A softwaer watchdog crash fixed.
* [FIX] Fixed a Lua -> Redis numerical precision loss.
* [NEW] Lua scripting engine speed improved.
* [NEW] Sentinel generates one new event for humans to understand better
what is happening during a failover: +config-update-from.
Also the time at which a failover will be re-attempted is logged.
--[ Redis 2.8.9 ] Release date: 22 Apr 2014
# UPGRADE URGENCY: LOW, only new features introduced, no bugs fixed.
* [NEW] The HyperLogLog data structure. You can read more about it
in this blog post. http://antirez.com/news/75
* [NEW] The Sorted Set data type has now support for lexicographic range
queries, check the new commands ZRANGEBYLEX, ZLEXCOUNT and
ZREMRANGEBYLEX, which are documented at http://redis.io.
--[ Redis 2.8.8 ] Release date: 25 Mar 2014
# UPGRADE URGENCY: HIGH for Redis, LOW for Sentinel. There is a potentially
critical bug fix causing data loss in Redis but it requires
a combination of disk full and the use of the
SHUTDOWN command.
* [FIX] Fixed data loss when SHUTDOWN was used with a disk full condition.
* [FIX] Fixed a memory leak in the SORT syntax error processing.
* [FIX] When Sentinel down-after-milliseconds parameter is modified at runtime
now it gets propagated to all the slaves and sentinel instances
of the master.
* [FIX] `install_server.sh` script finally fixed.
* [FIX] Different fixes to maxclients handling.
* [NEW] Sentinels are now able to send update messages in a peer-to-peer
fashion even if no Redis instances are available. Now the Sentinel
liveness property that the most updated configuration in a given
partition is propagated to all the Sentinels is extended to partitions
without reachable instances.
* [NEW] Sentinel safety properties are now ensured in a crash-recovery system
model since some state is persisted on disk before replying to other
nodes, and reloaded at startup.
* [NEW] Sentinel now uses CLIENT SETNAME so that it is easy to identify
Sentinels using CLIENT LIST among other clients.
* [NEW] Sentinel failure detection and reconnection code improved.
* [NEW] Use all 24 bits (instead of 22) for the Redis objects LRU field.
Note that the new LRU algorithm using eviction pools was not backported
from unstable for safery / code maturity concerns.
* [NEW] Majory speedup for the INFO command (it is now 6 times faster).
* [NEW] More Sentinel unit tests.
* [NEW] New command DEBUG ERROR returns the specified error. Example:
DEBUG ERROR "LOADING database". This is handy to write Redis client
libraries unit tests.
* [NEW] redis-cli now supports multi-line editing via updated linenoise lib.
Thanks to Matt Stancliff and Jan-Erik Rediger for the work done in the context
of this release.
--[ Redis 2.8.7 ] Release date: 5 Mar 2014
# UPGRADE URGENCY: LOW for Redis, LOW for Sentinel. However this release adds
@@ -269,6 +342,9 @@ that you should be aware of:
The following commands changed behavior:
* The TTL and PTTL commands now return -2 if the key does not exist and
-1 if it exists but has no associated expire. Redis 2.6 and previous
versions used to return -1 for both the conditions.
* SORT with ALPHA now sorts according to local collation locale if no STORE
option is used.
* ZADD/ZINCRBY are now able to accept a bigger range of values as valid
+1 -1
View File
@@ -130,7 +130,7 @@ it the proper way for a production system, we have a script doing this
for Ubuntu and Debian systems:
% cd utils
% ./install_server
% ./install_server.sh
The script will ask you a few questions and will setup everything you need
to run Redis properly as a background daemon that will start again on
+693 -229
View File
File diff suppressed because it is too large Load Diff
+34 -23
View File
@@ -3,39 +3,44 @@
*
* See linenoise.c for more information.
*
* ------------------------------------------------------------------------
*
* Copyright (c) 2010, Salvatore Sanfilippo <antirez at gmail dot com>
* Copyright (c) 2010, Pieter Noordhuis <pcnoordhuis at gmail dot com>
*
* All rights reserved.
*
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
* modification, are permitted provided that the following conditions are
* met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* * Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* * Neither the name of Redis nor the names of its contributors may be used
* to endorse or promote products derived from this software without
* specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __LINENOISE_H
#define __LINENOISE_H
#ifdef __cplusplus
extern "C" {
#endif
typedef struct linenoiseCompletions {
size_t len;
char **cvec;
@@ -43,13 +48,19 @@ typedef struct linenoiseCompletions {
typedef void(linenoiseCompletionCallback)(const char *, linenoiseCompletions *);
void linenoiseSetCompletionCallback(linenoiseCompletionCallback *);
void linenoiseAddCompletion(linenoiseCompletions *, char *);
void linenoiseAddCompletion(linenoiseCompletions *, const char *);
char *linenoise(const char *prompt);
int linenoiseHistoryAdd(const char *line);
int linenoiseHistorySetMaxLen(int len);
int linenoiseHistorySave(char *filename);
int linenoiseHistoryLoad(char *filename);
int linenoiseHistorySave(const char *filename);
int linenoiseHistoryLoad(const char *filename);
void linenoiseClearScreen(void);
void linenoiseSetMultiLine(int ml);
void linenoisePrintKeyCodes(void);
#ifdef __cplusplus
}
#endif
#endif /* __LINENOISE_H */
+14
View File
@@ -635,6 +635,20 @@ set-max-intset-entries 512
zset-max-ziplist-entries 128
zset-max-ziplist-value 64
# HyperLogLog sparse representation bytes limit. The limit includes the
# 16 bytes header. When an HyperLogLog using the sparse representation crosses
# this limit, it is converted into the dense representation.
#
# A value greater than 16000 is totally useless, since at that point the
# dense representation is more memory efficient.
#
# The suggested value is ~ 3000 in order to have the benefits of
# the space efficient encoding without slowing down too much PFADD,
# which is O(N) with the sparse encoding. The value can be raised to
# ~ 10000 when CPU is not a concern, but space is, and the data set is
# composed of many HyperLogLogs with cardinality in the 0 - 15000 range.
hll-sparse-max-bytes 3000
# Active rehashing uses 1 millisecond every 100 milliseconds of CPU time in
# order to help rehashing the main Redis hash table (the one mapping top-level
# keys to values). The hash table implementation Redis uses (see dict.c)
+15 -2
View File
@@ -4,6 +4,13 @@
# The port that this sentinel instance will run on
port 26379
# dir <working-directory>
# Every long running process should have a well-defined working directory.
# For Redis Sentinel to chdir to /tmp at startup is the simplest thing
# for the process to don't interferer with administrative tasks such as
# unmounting filesystems.
dir /tmp
# sentinel monitor <master-name> <ip> <redis-port> <quorum>
#
# Tells Sentinel to monitor this master, and to consider it in O_DOWN
@@ -13,6 +20,12 @@ port 26379
# be elected by the majority of the known Sentinels in order to
# start a failover, so no failover can be performed in minority.
#
# Slaves are auto-discovered, so you don't need to specify slaves in
# any way. Sentinel itself will rewrite this configuration file adding
# the slaves using additional configuration options.
# Also note that the configuration file is rewritten when a
# slave is promoted to master.
#
# Note: master name should not include special characters or spaces.
# The valid charset is A-z 0-9 and the three characters ".-_".
sentinel monitor mymaster 127.0.0.1 6379 2
@@ -86,10 +99,10 @@ sentinel failover-timeout mymaster 180000
# or to reconfigure clients after a failover. The scripts are executed
# with the following rules for error handling:
#
# If script exists with "1" the execution is retried later (up to a maximum
# If script exits with "1" the execution is retried later (up to a maximum
# number of times currently set to 10).
#
# If script exists with "2" (or an higher value) the script execution is
# If script exits with "2" (or an higher value) the script execution is
# not retried.
#
# If script terminates because it receives a signal the behavior is the same
+1 -1
View File
@@ -107,7 +107,7 @@ endif
REDIS_SERVER_NAME=redis-server
REDIS_SENTINEL_NAME=redis-sentinel
REDIS_SERVER_OBJ=adlist.o ae.o anet.o dict.o redis.o sds.o zmalloc.o lzf_c.o lzf_d.o pqsort.o zipmap.o sha1.o ziplist.o release.o networking.o util.o object.o db.o replication.o rdb.o t_string.o t_list.o t_set.o t_zset.o t_hash.o config.o aof.o pubsub.o multi.o debug.o sort.o intset.o syncio.o migrate.o endianconv.o slowlog.o scripting.o bio.o rio.o rand.o memtest.o crc64.o bitops.o sentinel.o notify.o setproctitle.o
REDIS_SERVER_OBJ=adlist.o ae.o anet.o dict.o redis.o sds.o zmalloc.o lzf_c.o lzf_d.o pqsort.o zipmap.o sha1.o ziplist.o release.o networking.o util.o object.o db.o replication.o rdb.o t_string.o t_list.o t_set.o t_zset.o t_hash.o config.o aof.o pubsub.o multi.o debug.o sort.o intset.o syncio.o migrate.o endianconv.o slowlog.o scripting.o bio.o rio.o rand.o memtest.o crc64.o bitops.o sentinel.o notify.o setproctitle.o hyperloglog.o
REDIS_CLI_NAME=redis-cli
REDIS_CLI_OBJ=anet.o sds.o adlist.o redis-cli.o zmalloc.o release.o anet.o ae.o crc64.o
REDIS_BENCHMARK_NAME=redis-benchmark
+4
View File
@@ -26,6 +26,10 @@ debug.o: debug.c redis.h fmacros.h config.h ../deps/lua/src/lua.h \
ziplist.h intset.h version.h util.h rdb.h rio.h sha1.h crc64.h bio.h
dict.o: dict.c fmacros.h dict.h zmalloc.h redisassert.h
endianconv.o: endianconv.c
hyperloglog.o: hyperloglog.c redis.h fmacros.h config.h \
../deps/lua/src/lua.h ../deps/lua/src/luaconf.h ae.h sds.h dict.h \
adlist.h zmalloc.h anet.h ziplist.h intset.h version.h util.h rdb.h \
rio.h
intset.o: intset.c intset.h zmalloc.h endianconv.h config.h
lzf_c.o: lzf_c.c lzfP.h
lzf_d.o: lzf_d.c lzfP.h
+2 -2
View File
@@ -472,7 +472,7 @@ int anetTcpAccept(char *err, int s, char *ip, size_t ip_len, int *port) {
int fd;
struct sockaddr_storage sa;
socklen_t salen = sizeof(sa);
if ((fd = anetGenericAccept(err,s,(struct sockaddr*)&sa,&salen)) == ANET_ERR)
if ((fd = anetGenericAccept(err,s,(struct sockaddr*)&sa,&salen)) == -1)
return ANET_ERR;
if (sa.ss_family == AF_INET) {
@@ -491,7 +491,7 @@ int anetUnixAccept(char *err, int s) {
int fd;
struct sockaddr_un sa;
socklen_t salen = sizeof(sa);
if ((fd = anetGenericAccept(err,s,(struct sockaddr*)&sa,&salen)) == ANET_ERR)
if ((fd = anetGenericAccept(err,s,(struct sockaddr*)&sa,&salen)) == -1)
return ANET_ERR;
return fd;
+4 -4
View File
@@ -559,7 +559,7 @@ int loadAppendOnlyFile(char *filename) {
/* Serve the clients from time to time */
if (!(loops++ % 1000)) {
loadingProgress(ftello(fp));
aeProcessEvents(server.el, AE_FILE_EVENTS|AE_DONT_WAIT);
processEventsWhileBlocked();
}
if (fgets(buf,sizeof(buf),fp) == NULL) {
@@ -968,9 +968,9 @@ int rewriteAppendOnlyFile(char *filename) {
}
/* Make sure data will not remain on the OS's output buffers */
fflush(fp);
aof_fsync(fileno(fp));
fclose(fp);
if (fflush(fp) == EOF) goto werr;
if (aof_fsync(fileno(fp)) == -1) goto werr;
if (fclose(fp) == EOF) goto werr;
/* Use RENAME to make sure the DB file is changed atomically only
* if the generate DB file is ok. */
+1 -8
View File
@@ -223,14 +223,7 @@ void setbitCommand(redisClient *c) {
dbAdd(c->db,c->argv[1],o);
} else {
if (checkType(c,o,REDIS_STRING)) return;
/* Create a copy when the object is shared or encoded. */
if (o->refcount != 1 || o->encoding != REDIS_ENCODING_RAW) {
robj *decoded = getDecodedObject(o);
o = createStringObject(decoded->ptr, sdslen(decoded->ptr));
decrRefCount(decoded);
dbOverwrite(c->db,c->argv[1],o);
}
o = dbUnshareStringValue(c->db,c->argv[1],o);
}
/* Grow sds value to the right length if necessary */
+10 -9
View File
@@ -388,6 +388,8 @@ void loadServerConfigFromString(char *config) {
server.zset_max_ziplist_entries = memtoll(argv[1], NULL);
} else if (!strcasecmp(argv[0],"zset-max-ziplist-value") && argc == 2) {
server.zset_max_ziplist_value = memtoll(argv[1], NULL);
} else if (!strcasecmp(argv[0],"hll-sparse-max-bytes") && argc == 2) {
server.hll_sparse_max_bytes = memtoll(argv[1], NULL);
} else if (!strcasecmp(argv[0],"rename-command") && argc == 3) {
struct redisCommand *cmd = lookupCommand(argv[1]);
int retval;
@@ -568,7 +570,7 @@ void configSetCommand(redisClient *c) {
} else if (!strcasecmp(c->argv[2]->ptr,"maxclients")) {
int orig_value = server.maxclients;
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt;
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 1) goto badfmt;
/* Try to check if the OS is capable of supporting so many FDs. */
server.maxclients = ll;
@@ -733,6 +735,9 @@ void configSetCommand(redisClient *c) {
} else if (!strcasecmp(c->argv[2]->ptr,"zset-max-ziplist-value")) {
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt;
server.zset_max_ziplist_value = ll;
} else if (!strcasecmp(c->argv[2]->ptr,"hll-sparse-max-bytes")) {
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt;
server.hll_sparse_max_bytes = ll;
} else if (!strcasecmp(c->argv[2]->ptr,"lua-time-limit")) {
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll < 0) goto badfmt;
server.lua_time_limit = ll;
@@ -934,6 +939,8 @@ void configGetCommand(redisClient *c) {
server.zset_max_ziplist_entries);
config_get_numerical_field("zset-max-ziplist-value",
server.zset_max_ziplist_value);
config_get_numerical_field("hll-sparse-max-bytes",
server.hll_sparse_max_bytes);
config_get_numerical_field("lua-time-limit",server.lua_time_limit);
config_get_numerical_field("slowlog-log-slower-than",
server.slowlog_log_slower_than);
@@ -1726,6 +1733,7 @@ int rewriteConfig(char *path) {
rewriteConfigNumericalOption(state,"set-max-intset-entries",server.set_max_intset_entries,REDIS_SET_MAX_INTSET_ENTRIES);
rewriteConfigNumericalOption(state,"zset-max-ziplist-entries",server.zset_max_ziplist_entries,REDIS_ZSET_MAX_ZIPLIST_ENTRIES);
rewriteConfigNumericalOption(state,"zset-max-ziplist-value",server.zset_max_ziplist_value,REDIS_ZSET_MAX_ZIPLIST_VALUE);
rewriteConfigNumericalOption(state,"hll-sparse-max-bytes",server.hll_sparse_max_bytes,REDIS_DEFAULT_HLL_SPARSE_MAX_BYTES);
rewriteConfigYesNoOption(state,"activerehashing",server.activerehashing,REDIS_DEFAULT_ACTIVE_REHASHING);
rewriteConfigClientoutputbufferlimitOption(state);
rewriteConfigNumericalOption(state,"hz",server.hz,REDIS_DEFAULT_HZ);
@@ -1760,14 +1768,7 @@ void configCommand(redisClient *c) {
configGetCommand(c);
} else if (!strcasecmp(c->argv[1]->ptr,"resetstat")) {
if (c->argc != 2) goto badarity;
server.stat_keyspace_hits = 0;
server.stat_keyspace_misses = 0;
server.stat_numcommands = 0;
server.stat_numconnections = 0;
server.stat_expiredkeys = 0;
server.stat_rejected_conn = 0;
server.stat_fork_time = 0;
server.aof_delayed_fsync = 0;
resetServerStats();
resetCommandTableStats();
addReply(c,shared.ok);
} else if (!strcasecmp(c->argv[1]->ptr,"rewrite")) {
+1 -1
View File
@@ -187,7 +187,7 @@ void setproctitle(const char *fmt, ...);
#if (__i386 || __amd64) && __GNUC__
#define GNUC_VERSION (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__)
#if GNUC_VERSION >= 40100
#if (GNUC_VERSION >= 40100) || defined(__clang__)
#define HAVE_ATOMIC
#endif
#endif
+39
View File
@@ -166,6 +166,44 @@ int dbDelete(redisDb *db, robj *key) {
}
}
/* Prepare the string object stored at 'key' to be modified destructively
* to implement commands like SETBIT or APPEND.
*
* An object is usually ready to be modified unless one of the two conditions
* are true:
*
* 1) The object 'o' is shared (refcount > 1), we don't want to affect
* other users.
* 2) The object encoding is not "RAW".
*
* If the object is found in one of the above conditions (or both) by the
* function, an unshared / not-encoded copy of the string object is stored
* at 'key' in the specified 'db'. Otherwise the object 'o' itself is
* returned.
*
* USAGE:
*
* The object 'o' is what the caller already obtained by looking up 'key'
* in 'db', the usage pattern looks like this:
*
* o = lookupKeyWrite(db,key);
* if (checkType(c,o,REDIS_STRING)) return;
* o = dbUnshareStringValue(db,key,o);
*
* At this point the caller is ready to modify the object, for example
* using an sdscat() call to append some data, or anything else.
*/
robj *dbUnshareStringValue(redisDb *db, robj *key, robj *o) {
redisAssert(o->type == REDIS_STRING);
if (o->refcount != 1 || o->encoding != REDIS_ENCODING_RAW) {
robj *decoded = getDecodedObject(o);
o = createStringObject(decoded->ptr, sdslen(decoded->ptr));
decrRefCount(decoded);
dbOverwrite(db,key,o);
}
return o;
}
long long emptyDb(void(callback)(void*)) {
int j;
long long removed = 0;
@@ -236,6 +274,7 @@ void delCommand(redisClient *c) {
int deleted = 0, j;
for (j = 1; j < c->argc; j++) {
expireIfNeeded(c->db,c->argv[j]);
if (dbDelete(c->db,c->argv[j])) {
signalModifiedKey(c->db,c->argv[j]);
notifyKeyspaceEvent(REDIS_NOTIFY_GENERIC,
+9 -1
View File
@@ -326,6 +326,7 @@ void debugCommand(redisClient *c) {
if (getLongFromObjectOrReply(c, c->argv[2], &keys, NULL) != REDIS_OK)
return;
dictExpand(c->db->dict,keys);
for (j = 0; j < keys; j++) {
snprintf(buf,sizeof(buf),"key:%lu",j);
key = createStringObject(buf,strlen(buf));
@@ -363,6 +364,13 @@ void debugCommand(redisClient *c) {
{
server.active_expire_enabled = atoi(c->argv[2]->ptr);
addReply(c,shared.ok);
} else if (!strcasecmp(c->argv[1]->ptr,"error") && c->argc == 3) {
sds errstr = sdsnewlen("-",1);
errstr = sdscatsds(errstr,c->argv[2]->ptr);
errstr = sdsmapchars(errstr,"\n\r"," ",2); /* no newlines in errors. */
errstr = sdscatlen(errstr,"\r\n",2);
addReplySds(c,errstr);
} else {
addReplyErrorFormat(c, "Unknown DEBUG subcommand or wrong number of arguments for '%s'",
(char*)c->argv[1]->ptr);
@@ -923,7 +931,7 @@ void enableWatchdog(int period) {
/* Watchdog was actually disabled, so we have to setup the signal
* handler. */
sigemptyset(&act.sa_mask);
act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_SIGINFO;
act.sa_flags = SA_ONSTACK | SA_SIGINFO;
act.sa_sigaction = watchdogSignalHandler;
sigaction(SIGALRM, &act, NULL);
}
+2 -2
View File
@@ -239,7 +239,7 @@ int dictExpand(dict *d, unsigned long size)
/* Performs N steps of incremental rehashing. Returns 1 if there are still
* keys to move from the old to the new hash table, otherwise 0 is returned.
* Note that a rehashing step consists in moving a bucket (that may have more
* thank one key as we use chaining) from the old to the new hash table. */
* than one key as we use chaining) from the old to the new hash table. */
int dictRehash(dict *d, int n) {
if (!dictIsRehashing(d)) return 0;
@@ -695,7 +695,7 @@ static unsigned long rev(unsigned long v) {
* (where SIZE-1 is always the mask that is equivalent to taking the rest
* of the division between the Hash of the key and SIZE).
*
* For example if the current hash table size is 64, the mask is
* For example if the current hash table size is 16, the mask is
* (in binary) 1111. The position of a key in the hash table will be always
* the last four bits of the hash output, and so forth.
*
+73 -2
View File
@@ -14,7 +14,8 @@ static char *commandGroups[] = {
"transactions",
"connection",
"server",
"scripting"
"scripting",
"hyperloglog"
};
struct commandHelp {
@@ -54,6 +55,11 @@ struct commandHelp {
"Perform bitwise operations between strings",
1,
"2.6.0" },
{ "BITPOS",
"key bit [start] [end]",
"Find first bit set or clear in a string",
1,
"2.8.7" },
{ "BLPOP",
"key [key ...] timeout",
"Remove and get the first element in a list, or block until one is available",
@@ -84,6 +90,11 @@ struct commandHelp {
"Get the list of client connections",
9,
"2.4.0" },
{ "CLIENT PAUSE",
"timeout",
"Stop processing commands from clients for some time",
9,
"2.9.50" },
{ "CLIENT SETNAME",
"connection-name",
"Set the current connection name",
@@ -99,6 +110,11 @@ struct commandHelp {
"Reset the stats returned by INFO",
9,
"2.0.0" },
{ "CONFIG REWRITE",
"-",
"Rewrite the configuration file with the in memory configuration",
9,
"2.8.0" },
{ "CONFIG SET",
"parameter value",
"Set a configuration parameter to the given value",
@@ -259,6 +275,11 @@ struct commandHelp {
"Set multiple hash fields to multiple values",
5,
"2.0.0" },
{ "HSCAN",
"key cursor [MATCH pattern] [COUNT count]",
"Incrementally iterate hash fields and associated values",
5,
"2.8.0" },
{ "HSET",
"key field value",
"Set the string value of a hash field",
@@ -360,7 +381,7 @@ struct commandHelp {
1,
"1.0.0" },
{ "MIGRATE",
"host port key destination-db timeout",
"host port key destination-db timeout [COPY] [REPLACE]",
"Atomically transfer a key from a Redis instance to another one.",
0,
"2.6.0" },
@@ -409,6 +430,21 @@ struct commandHelp {
"Set the expiration for a key as a UNIX timestamp specified in milliseconds",
0,
"2.6.0" },
{ "PFADD",
"key element [element ...]",
"Adds the specified elements to the specified HyperLogLog.",
11,
"2.8.9" },
{ "PFCOUNT",
"key [key ...]",
"Return the approximated cardinality of the set(s) observed by the HyperLogLog at key(s).",
11,
"2.8.9" },
{ "PFMERGE",
"destkey sourcekey [sourcekey ...]",
"Merge N different HyperLogLogs into a single one.",
11,
"2.8.9" },
{ "PING",
"-",
"Ping the server",
@@ -434,6 +470,11 @@ struct commandHelp {
"Post a message to a channel",
6,
"2.0.0" },
{ "PUBSUB",
"subcommand [argument [argument ...]]",
"Inspect the state of the Pub/Sub subsystem",
6,
"2.8.0" },
{ "PUNSUBSCRIBE",
"[pattern [pattern ...]]",
"Stop listening for messages posted to channels matching the given patterns",
@@ -494,6 +535,11 @@ struct commandHelp {
"Synchronously save the dataset to disk",
9,
"1.0.0" },
{ "SCAN",
"cursor [MATCH pattern] [COUNT count]",
"Incrementally iterate the keys space",
0,
"2.8.0" },
{ "SCARD",
"key",
"Get the number of members in a set",
@@ -619,6 +665,11 @@ struct commandHelp {
"Remove one or more members from a set",
3,
"1.0.0" },
{ "SSCAN",
"key cursor [MATCH pattern] [COUNT count]",
"Incrementally iterate Set elements",
3,
"2.8.0" },
{ "STRLEN",
"key",
"Get the length of the value stored in a key",
@@ -699,11 +750,21 @@ struct commandHelp {
"Intersect multiple sorted sets and store the resulting sorted set in a new key",
4,
"2.0.0" },
{ "ZLEXCOUNT",
"key min max",
"Count the number of members in a sorted set between a given lexicographical range",
4,
"2.8.9" },
{ "ZRANGE",
"key start stop [WITHSCORES]",
"Return a range of members in a sorted set, by index",
4,
"1.2.0" },
{ "ZRANGEBYLEX",
"key min max [LIMIT offset count]",
"Return a range of members in a sorted set, by lexicographical range",
4,
"2.8.9" },
{ "ZRANGEBYSCORE",
"key min max [WITHSCORES] [LIMIT offset count]",
"Return a range of members in a sorted set, by score",
@@ -719,6 +780,11 @@ struct commandHelp {
"Remove one or more members from a sorted set",
4,
"1.2.0" },
{ "ZREMRANGEBYLEX",
"key min max",
"Remove all members in a sorted set between the given lexicographical range",
4,
"2.8.9" },
{ "ZREMRANGEBYRANK",
"key start stop",
"Remove all members in a sorted set within the given indexes",
@@ -744,6 +810,11 @@ struct commandHelp {
"Determine the index of a member in a sorted set, with scores ordered from high to low",
4,
"2.0.0" },
{ "ZSCAN",
"key cursor [MATCH pattern] [COUNT count]",
"Incrementally iterate sorted sets elements and associated scores",
4,
"2.8.0" },
{ "ZSCORE",
"key member",
"Get the score associated with the given member in a sorted set",
+1548
View File
File diff suppressed because it is too large Load Diff
+49 -16
View File
@@ -534,6 +534,7 @@ void copyClientOutputBuffer(redisClient *dst, redisClient *src) {
dst->reply_bytes = src->reply_bytes;
}
#define MAX_ACCEPTS_PER_CALL 1000
static void acceptCommonHandler(int fd, int flags) {
redisClient *c;
if ((c = createClient(fd)) == NULL) {
@@ -563,37 +564,44 @@ static void acceptCommonHandler(int fd, int flags) {
}
void acceptTcpHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
int cport, cfd;
int cport, cfd, max = MAX_ACCEPTS_PER_CALL;
char cip[REDIS_IP_STR_LEN];
REDIS_NOTUSED(el);
REDIS_NOTUSED(mask);
REDIS_NOTUSED(privdata);
cfd = anetTcpAccept(server.neterr, fd, cip, sizeof(cip), &cport);
if (cfd == AE_ERR) {
redisLog(REDIS_WARNING,"Accepting client connection: %s", server.neterr);
return;
while(max--) {
cfd = anetTcpAccept(server.neterr, fd, cip, sizeof(cip), &cport);
if (cfd == ANET_ERR) {
if (errno != EWOULDBLOCK)
redisLog(REDIS_WARNING,
"Accepting client connection: %s", server.neterr);
return;
}
redisLog(REDIS_VERBOSE,"Accepted %s:%d", cip, cport);
acceptCommonHandler(cfd,0);
}
redisLog(REDIS_VERBOSE,"Accepted %s:%d", cip, cport);
acceptCommonHandler(cfd,0);
}
void acceptUnixHandler(aeEventLoop *el, int fd, void *privdata, int mask) {
int cfd;
int cfd, max = MAX_ACCEPTS_PER_CALL;
REDIS_NOTUSED(el);
REDIS_NOTUSED(mask);
REDIS_NOTUSED(privdata);
cfd = anetUnixAccept(server.neterr, fd);
if (cfd == AE_ERR) {
redisLog(REDIS_WARNING,"Accepting client connection: %s", server.neterr);
return;
while(max--) {
cfd = anetUnixAccept(server.neterr, fd);
if (cfd == ANET_ERR) {
if (errno != EWOULDBLOCK)
redisLog(REDIS_WARNING,
"Accepting client connection: %s", server.neterr);
return;
}
redisLog(REDIS_VERBOSE,"Accepted connection to %s", server.unixsocket);
acceptCommonHandler(cfd,REDIS_UNIX_SOCKET);
}
redisLog(REDIS_VERBOSE,"Accepted connection to %s", server.unixsocket);
acceptCommonHandler(cfd,REDIS_UNIX_SOCKET);
}
static void freeClientArgv(redisClient *c) {
int j;
for (j = 0; j < c->argc; j++)
@@ -977,8 +985,10 @@ int processMultibulkBuffer(redisClient *c) {
newline = strchr(c->querybuf+pos,'\r');
if (newline == NULL) {
if (sdslen(c->querybuf) > REDIS_INLINE_MAX_SIZE) {
addReplyError(c,"Protocol error: too big bulk count string");
addReplyError(c,
"Protocol error: too big bulk count string");
setProtocolError(c,0);
return REDIS_ERR;
}
break;
}
@@ -1531,3 +1541,26 @@ void flushSlavesOutputBuffers(void) {
}
}
}
/* This function is called by Redis in order to process a few events from
* time to time while blocked into some not interruptible operation.
* This allows to reply to clients with the -LOADING error while loading the
* data set at startup or after a full resynchronization with the master
* and so forth.
*
* It calls the event loop in order to process a few events. Specifically we
* try to call the event loop for times as long as we receive acknowledge that
* some event was processed, in order to go forward with the accept, read,
* write, close sequence needed to serve a client.
*
* The function returns the total number of events processed. */
int processEventsWhileBlocked(void) {
int iterations = 4; /* See the function top-comment. */
int count = 0;
while (iterations--) {
int events = aeProcessEvents(server.el, AE_FILE_EVENTS|AE_DONT_WAIT);
if (!events) break;
count += events;
}
return count;
}
+1 -1
View File
@@ -587,7 +587,7 @@ unsigned long estimateObjectIdleTime(robj *o) {
}
}
/* This is a helper function for the DEBUG command. We need to lookup keys
/* This is a helper function for the OBJECT command. We need to lookup keys
* without any modification of LRU or other parameters. */
robj *objectCommandLookup(redisClient *c, robj *key) {
dictEntry *de;
+6 -6
View File
@@ -398,7 +398,7 @@ int rdbSaveDoubleValue(rio *rdb, double val) {
double min = -4503599627370495; /* (2^52)-1 */
double max = 4503599627370496; /* -(2^52) */
if (val > min && val < max && val == ((double)((long long)val)))
ll2string((char*)buf+1,sizeof(buf),(long long)val);
ll2string((char*)buf+1,sizeof(buf)-1,(long long)val);
else
#endif
snprintf((char*)buf+1,sizeof(buf)-1,"%.17g",val);
@@ -410,7 +410,7 @@ int rdbSaveDoubleValue(rio *rdb, double val) {
/* For information about double serialization check rdbSaveDoubleValue() */
int rdbLoadDoubleValue(rio *rdb, double *val) {
char buf[128];
char buf[256];
unsigned char len;
if (rioRead(rdb,&len,1) == 0) return -1;
@@ -690,9 +690,9 @@ int rdbSave(char *filename) {
rioWrite(&rdb,&cksum,8);
/* Make sure data will not remain on the OS's output buffers */
fflush(fp);
fsync(fileno(fp));
fclose(fp);
if (fflush(fp) == EOF) goto werr;
if (fsync(fileno(fp)) == -1) goto werr;
if (fclose(fp) == EOF) goto werr;
/* Use RENAME to make sure the DB file is changed atomically only
* if the generate DB file is ok. */
@@ -1073,7 +1073,7 @@ void rdbLoadProgressCallback(rio *r, const void *buf, size_t len) {
if (server.masterhost && server.repl_state == REDIS_REPL_TRANSFER)
replicationSendNewlineToMaster();
loadingProgress(r->processed_bytes);
aeProcessEvents(server.el, AE_FILE_EVENTS|AE_DONT_WAIT);
processEventsWhileBlocked();
}
}
+1
View File
@@ -865,6 +865,7 @@ static void repl() {
sds *argv;
config.interactive = 1;
linenoiseSetMultiLine(1);
linenoiseSetCompletionCallback(completionCallback);
/* Only use history when stdin is a tty. */
+115 -35
View File
@@ -168,12 +168,16 @@ struct redisCommand redisCommandTable[] = {
{"zrem",zremCommand,-3,"w",0,NULL,1,1,1,0,0},
{"zremrangebyscore",zremrangebyscoreCommand,4,"w",0,NULL,1,1,1,0,0},
{"zremrangebyrank",zremrangebyrankCommand,4,"w",0,NULL,1,1,1,0,0},
{"zremrangebylex",zremrangebylexCommand,4,"w",0,NULL,1,1,1,0,0},
{"zunionstore",zunionstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
{"zinterstore",zinterstoreCommand,-4,"wm",0,zunionInterGetKeys,0,0,0,0,0},
{"zrange",zrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
{"zrangebyscore",zrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
{"zrevrangebyscore",zrevrangebyscoreCommand,-4,"r",0,NULL,1,1,1,0,0},
{"zrangebylex",zrangebylexCommand,-4,"r",0,NULL,1,1,1,0,0},
{"zrevrangebylex",zrevrangebylexCommand,-4,"r",0,NULL,1,1,1,0,0},
{"zcount",zcountCommand,4,"r",0,NULL,1,1,1,0,0},
{"zlexcount",zlexcountCommand,4,"r",0,NULL,1,1,1,0,0},
{"zrevrange",zrevrangeCommand,-4,"r",0,NULL,1,1,1,0,0},
{"zcard",zcardCommand,2,"r",0,NULL,1,1,1,0,0},
{"zscore",zscoreCommand,3,"r",0,NULL,1,1,1,0,0},
@@ -258,7 +262,12 @@ struct redisCommand redisCommandTable[] = {
{"time",timeCommand,1,"rR",0,NULL,0,0,0,0,0},
{"bitop",bitopCommand,-4,"wm",0,NULL,2,-1,1,0,0},
{"bitcount",bitcountCommand,-2,"r",0,NULL,1,1,1,0,0},
{"bitpos",bitposCommand,-3,"r",0,NULL,1,1,1,0,0}
{"bitpos",bitposCommand,-3,"r",0,NULL,1,1,1,0,0},
{"pfselftest",pfselftestCommand,1,"r",0,NULL,0,0,0,0,0},
{"pfadd",pfaddCommand,-2,"wm",0,NULL,1,1,1,0,0},
{"pfcount",pfcountCommand,-2,"w",0,NULL,1,1,1,0,0},
{"pfmerge",pfmergeCommand,-2,"wm",0,NULL,1,-1,1,0,0},
{"pfdebug",pfdebugCommand,-3,"w",0,NULL,0,0,0,0,0}
};
/*============================ Utility functions ============================ */
@@ -1016,6 +1025,9 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
if (zmalloc_used_memory() > server.stat_peak_memory)
server.stat_peak_memory = zmalloc_used_memory();
/* Sample the RSS here since this is a relatively slow call. */
server.resident_set_size = zmalloc_get_rss();
/* We received a SIGTERM, shutting down here in a safe way, as it is
* not ok doing so inside the signal handler. */
if (server.shutdown_asap) {
@@ -1270,6 +1282,12 @@ void createSharedObjects(void) {
shared.bulkhdr[j] = createObject(REDIS_STRING,
sdscatprintf(sdsempty(),"$%d\r\n",j));
}
/* The following two shared objects, minstring and maxstrings, are not
* actually used for their value but as a special object meaning
* respectively the minimum possible string and the maximum possible
* string in string comparisons for the ZRANGEBYLEX command. */
shared.minstring = createStringObject("minstring",9);
shared.maxstring = createStringObject("maxstring",9);
}
void initServerConfig() {
@@ -1337,6 +1355,7 @@ void initServerConfig() {
server.set_max_intset_entries = REDIS_SET_MAX_INTSET_ENTRIES;
server.zset_max_ziplist_entries = REDIS_ZSET_MAX_ZIPLIST_ENTRIES;
server.zset_max_ziplist_value = REDIS_ZSET_MAX_ZIPLIST_VALUE;
server.hll_sparse_max_bytes = REDIS_DEFAULT_HLL_SPARSE_MAX_BYTES;
server.shutdown_asap = 0;
server.repl_ping_slave_period = REDIS_REPL_PING_SLAVE_PERIOD;
server.repl_timeout = REDIS_REPL_TIMEOUT;
@@ -1414,21 +1433,21 @@ void initServerConfig() {
}
/* This function will try to raise the max number of open files accordingly to
* the configured max number of clients. It will also account for 32 additional
* file descriptors as we need a few more for persistence, listening
* sockets, log files and so forth.
* the configured max number of clients. It also reserves a number of file
* descriptors (REDIS_MIN_RESERVED_FDS) for extra operations of
* persistence, listening sockets, log files and so forth.
*
* If it will not be possible to set the limit accordingly to the configured
* max number of clients, the function will do the reverse setting
* server.maxclients to the value that we can actually handle. */
void adjustOpenFilesLimit(void) {
rlim_t maxfiles = server.maxclients+32;
rlim_t maxfiles = server.maxclients+REDIS_MIN_RESERVED_FDS;
struct rlimit limit;
if (getrlimit(RLIMIT_NOFILE,&limit) == -1) {
redisLog(REDIS_WARNING,"Unable to obtain the current NOFILE limit (%s), assuming 1024 and setting the max clients configuration accordingly.",
strerror(errno));
server.maxclients = 1024-32;
server.maxclients = 1024-REDIS_MIN_RESERVED_FDS;
} else {
rlim_t oldlimit = limit.rlim_cur;
@@ -1436,22 +1455,58 @@ void adjustOpenFilesLimit(void) {
* for our needs. */
if (oldlimit < maxfiles) {
rlim_t f;
int setrlimit_error = 0;
/* Try to set the file limit to match 'maxfiles' or at least
* to the higher value supported less than maxfiles. */
f = maxfiles;
while(f > oldlimit) {
int decr_step = 16;
limit.rlim_cur = f;
limit.rlim_max = f;
if (setrlimit(RLIMIT_NOFILE,&limit) != -1) break;
f -= 128;
setrlimit_error = errno;
/* We failed to set file limit to 'f'. Try with a
* smaller limit decrementing by a few FDs per iteration. */
if (f < decr_step) break;
f -= decr_step;
}
/* Assume that the limit we get initially is still valid if
* our last try was even lower. */
if (f < oldlimit) f = oldlimit;
if (f != maxfiles) {
server.maxclients = f-32;
redisLog(REDIS_WARNING,"Unable to set the max number of files limit to %d (%s), setting the max clients configuration to %d.",
(int) maxfiles, strerror(errno), (int) server.maxclients);
int old_maxclients = server.maxclients;
server.maxclients = f-REDIS_MIN_RESERVED_FDS;
if (server.maxclients < 1) {
redisLog(REDIS_WARNING,"Your current 'ulimit -n' "
"of %llu is not enough for Redis to start. "
"Please increase your open file limit to at least "
"%llu. Exiting.",
(unsigned long long) oldlimit,
(unsigned long long) maxfiles);
exit(1);
}
redisLog(REDIS_WARNING,"You requested maxclients of %d "
"requiring at least %llu max file descriptors.",
old_maxclients,
(unsigned long long) maxfiles);
redisLog(REDIS_WARNING,"Redis can't set maximum open files "
"to %llu because of OS error: %s.",
(unsigned long long) maxfiles, strerror(setrlimit_error));
redisLog(REDIS_WARNING,"Current maximum open files is %llu. "
"maxclients has been reduced to %d to compensate for "
"low ulimit. "
"If you need higher maxclients increase 'ulimit -n'.",
(unsigned long long) oldlimit, server.maxclients);
} else {
redisLog(REDIS_NOTICE,"Max number of open files set to %d",
(int) maxfiles);
redisLog(REDIS_NOTICE,"Increased maximum number of open files "
"to %llu (it was originally set to %llu).",
(unsigned long long) maxfiles,
(unsigned long long) oldlimit);
}
}
}
@@ -1487,10 +1542,16 @@ int listenToPort(int port, int *fds, int *count) {
* server.bindaddr_count == 0. */
fds[*count] = anetTcp6Server(server.neterr,port,NULL,
server.tcp_backlog);
if (fds[*count] != ANET_ERR) (*count)++;
if (fds[*count] != ANET_ERR) {
anetNonBlock(NULL,fds[*count]);
(*count)++;
}
fds[*count] = anetTcpServer(server.neterr,port,NULL,
server.tcp_backlog);
if (fds[*count] != ANET_ERR) (*count)++;
if (fds[*count] != ANET_ERR) {
anetNonBlock(NULL,fds[*count]);
(*count)++;
}
/* Exit the loop if we were able to bind * on IPv4 or IPv6,
* otherwise fds[*count] will be ANET_ERR and we'll print an
* error and return to the caller with an error. */
@@ -1511,11 +1572,33 @@ int listenToPort(int port, int *fds, int *count) {
server.port, server.neterr);
return REDIS_ERR;
}
anetNonBlock(NULL,fds[*count]);
(*count)++;
}
return REDIS_OK;
}
/* Resets the stats that we expose via INFO or other means that we want
* to reset via CONFIG RESETSTAT. The function is also used in order to
* initialize these fields in initServer() at server startup. */
void resetServerStats(void) {
server.stat_numcommands = 0;
server.stat_numconnections = 0;
server.stat_expiredkeys = 0;
server.stat_evictedkeys = 0;
server.stat_keyspace_misses = 0;
server.stat_keyspace_hits = 0;
server.stat_fork_time = 0;
server.stat_rejected_conn = 0;
server.stat_sync_full = 0;
server.stat_sync_partial_ok = 0;
server.stat_sync_partial_err = 0;
memset(server.ops_sec_samples,0,sizeof(server.ops_sec_samples));
server.ops_sec_idx = 0;
server.ops_sec_last_sample_time = mstime();
server.ops_sec_last_sample_ops = 0;
}
void initServer() {
int j;
@@ -1556,6 +1639,7 @@ void initServer() {
redisLog(REDIS_WARNING, "Opening socket: %s", server.neterr);
exit(1);
}
anetNonBlock(NULL,server.sofd);
}
/* Abort if there are no listening sockets at all. */
@@ -1588,23 +1672,11 @@ void initServer() {
server.rdb_save_time_last = -1;
server.rdb_save_time_start = -1;
server.dirty = 0;
server.stat_numcommands = 0;
server.stat_numconnections = 0;
server.stat_expiredkeys = 0;
server.stat_evictedkeys = 0;
resetServerStats();
/* A few stats we don't want to reset: server startup time, and peak mem. */
server.stat_starttime = time(NULL);
server.stat_keyspace_misses = 0;
server.stat_keyspace_hits = 0;
server.stat_peak_memory = 0;
server.stat_fork_time = 0;
server.stat_rejected_conn = 0;
server.stat_sync_full = 0;
server.stat_sync_partial_ok = 0;
server.stat_sync_partial_err = 0;
memset(server.ops_sec_samples,0,sizeof(server.ops_sec_samples));
server.ops_sec_idx = 0;
server.ops_sec_last_sample_time = mstime();
server.ops_sec_last_sample_ops = 0;
server.resident_set_size = 0;
server.lastbgsave_status = REDIS_OK;
server.aof_last_write_status = REDIS_OK;
server.aof_last_write_errno = 0;
@@ -1965,7 +2037,8 @@ int processCommand(redisClient *c) {
/* Don't accept write commands if there are not enough good slaves and
* user configured the min-slaves-to-write option. */
if (server.repl_min_slaves_to_write &&
if (server.masterhost == NULL &&
server.repl_min_slaves_to_write &&
server.repl_min_slaves_max_lag &&
c->cmd->flags & REDIS_CMD_WRITE &&
server.repl_good_slaves_count < server.repl_min_slaves_to_write)
@@ -2228,14 +2301,21 @@ sds genRedisInfoString(char *section) {
/* Server */
if (allsections || defsections || !strcasecmp(section,"server")) {
struct utsname name;
static int call_uname = 1;
static struct utsname name;
char *mode;
if (server.sentinel_mode) mode = "sentinel";
else mode = "standalone";
if (sections++) info = sdscat(info,"\r\n");
uname(&name);
if (call_uname) {
/* Uname can be slow and is always the same output. Cache it. */
uname(&name);
call_uname = 0;
}
info = sdscatprintf(info,
"# Server\r\n"
"redis_version:%s\r\n"
@@ -2320,11 +2400,11 @@ sds genRedisInfoString(char *section) {
"mem_allocator:%s\r\n",
zmalloc_used,
hmem,
zmalloc_get_rss(),
server.resident_set_size,
server.stat_peak_memory,
peak_hmem,
((long long)lua_gc(server.lua,LUA_GCCOUNT,0))*1024LL,
zmalloc_get_fragmentation_ratio(),
zmalloc_get_fragmentation_ratio(server.resident_set_size),
ZMALLOC_LIB
);
}
+36 -13
View File
@@ -123,6 +123,7 @@
#define REDIS_IP_STR_LEN INET6_ADDRSTRLEN
#define REDIS_PEER_ID_LEN (REDIS_IP_STR_LEN+32) /* Must be enough for ip:port */
#define REDIS_BINDADDR_MAX 16
#define REDIS_MIN_RESERVED_FDS 32
#define ACTIVE_EXPIRE_CYCLE_LOOKUPS_PER_LOOP 20 /* Loopkups per loop. */
#define ACTIVE_EXPIRE_CYCLE_FAST_DURATION 1000 /* Microseconds */
@@ -139,9 +140,9 @@
#define REDIS_LONGSTR_SIZE 21 /* Bytes needed for long -> str */
#define REDIS_AOF_AUTOSYNC_BYTES (1024*1024*32) /* fdatasync every 32MB */
/* When configuring the Redis eventloop, we setup it so that the total number
* of file descriptors we can handle are server.maxclients + FDSET_INCR
* of file descriptors we can handle are server.maxclients + RESERVED_FDS + FDSET_INCR
* that is our safety margin. */
#define REDIS_EVENTLOOP_FDSET_INCR 128
#define REDIS_EVENTLOOP_FDSET_INCR (REDIS_MIN_RESERVED_FDS+96)
/* Hash table parameters */
#define REDIS_HT_MINFILL 10 /* Minimal hash table fill 10% */
@@ -302,6 +303,9 @@
#define REDIS_ZSET_MAX_ZIPLIST_ENTRIES 128
#define REDIS_ZSET_MAX_ZIPLIST_VALUE 64
/* HyperLogLog defines */
#define REDIS_DEFAULT_HLL_SPARSE_MAX_BYTES 3000
/* Sets operations codes */
#define REDIS_OP_UNION 0
#define REDIS_OP_DIFF 1
@@ -373,13 +377,13 @@ typedef long long mstime_t; /* millisecond time type. */
/* A redis object, that is a type able to hold a string / list / set */
/* The actual Redis Object */
#define REDIS_LRU_CLOCK_MAX ((1<<21)-1) /* Max value of obj->lru */
#define REDIS_LRU_CLOCK_RESOLUTION 10 /* LRU clock resolution in seconds */
#define REDIS_LRU_BITS 24
#define REDIS_LRU_CLOCK_MAX ((1<<REDIS_LRU_BITS)-1) /* Max value of obj->lru */
#define REDIS_LRU_CLOCK_RESOLUTION 1 /* LRU clock resolution in seconds */
typedef struct redisObject {
unsigned type:4;
unsigned notused:2; /* Not used */
unsigned encoding:4;
unsigned lru:22; /* lru time (relative to server.lruclock) */
unsigned lru:REDIS_LRU_BITS; /* lru time (relative to server.lruclock) */
int refcount;
void *ptr;
} robj;
@@ -501,7 +505,7 @@ struct sharedObjectsStruct {
*masterdownerr, *roslaveerr, *execaborterr, *noautherr, *noreplicaserr,
*oomerr, *plus, *messagebulk, *pmessagebulk, *subscribebulk,
*unsubscribebulk, *psubscribebulk, *punsubscribebulk, *del, *rpop, *lpop,
*lpush, *emptyscan,
*lpush, *emptyscan, *minstring, *maxstring,
*select[REDIS_SHARED_SELECT_CMDS],
*integers[REDIS_SHARED_INTEGERS],
*mbulkhdr[REDIS_SHARED_BULKHDR_LEN], /* "*<value>\r\n" */
@@ -574,8 +578,7 @@ struct redisServer {
dict *commands; /* Command table */
dict *orig_commands; /* Command table before command renaming. */
aeEventLoop *el;
unsigned lruclock:22; /* Clock incrementing every minute, for LRU */
unsigned lruclock_padding:10;
unsigned lruclock:REDIS_LRU_BITS; /* Clock for LRU eviction */
int shutdown_asap; /* SHUTDOWN needed ASAP */
int activerehashing; /* Incremental rehash in serverCron() */
char *requirepass; /* Pass for AUTH command, or NULL */
@@ -626,6 +629,7 @@ struct redisServer {
long long slowlog_entry_id; /* SLOWLOG current entry ID */
long long slowlog_log_slower_than; /* SLOWLOG time limit (to get logged) */
unsigned long slowlog_max_len; /* SLOWLOG max number of items logged */
size_t resident_set_size; /* RSS sampled in serverCron(). */
/* The following two are used to track instantaneous "load" in terms
* of operations per second. */
long long ops_sec_last_sample_time; /* Timestamp of last sample (in ms) */
@@ -732,7 +736,7 @@ struct redisServer {
list *repl_scriptcache_fifo; /* First in, first out LRU eviction. */
int repl_scriptcache_size; /* Max number of elements. */
/* Limits */
unsigned int maxclients; /* Max number of simultaneous clients */
int maxclients; /* Max number of simultaneous clients */
unsigned long long maxmemory; /* Max number of memory bytes to use */
int maxmemory_policy; /* Policy for key eviction */
int maxmemory_samples; /* Pricision of random sampling */
@@ -754,6 +758,7 @@ struct redisServer {
size_t set_max_intset_entries;
size_t zset_max_ziplist_entries;
size_t zset_max_ziplist_value;
size_t hll_sparse_max_bytes;
time_t unixtime; /* Unix time sampled every cron cycle. */
long long mstime; /* Like 'unixtime' but with milliseconds resolution. */
/* Pubsub */
@@ -935,6 +940,7 @@ int getClientLimitClassByName(char *name);
char *getClientLimitClassName(int class);
void flushSlavesOutputBuffers(void);
void disconnectSlaves(void);
int processEventsWhileBlocked(void);
#ifdef __GNUC__
void addReplyErrorFormat(redisClient *c, const char *fmt, ...)
@@ -1056,18 +1062,24 @@ unsigned long aofRewriteBufferSize(void);
/* Sorted sets data type */
/* Struct to hold a inclusive/exclusive range spec. */
/* Struct to hold a inclusive/exclusive range spec by score comparison. */
typedef struct {
double min, max;
int minex, maxex; /* are min or max exclusive? */
} zrangespec;
/* Struct to hold an inclusive/exclusive range spec by lexicographic comparison. */
typedef struct {
robj *min, *max; /* May be set to shared.(minstring|maxstring) */
int minex, maxex; /* are min or max exclusive? */
} zlexrangespec;
zskiplist *zslCreate(void);
void zslFree(zskiplist *zsl);
zskiplistNode *zslInsert(zskiplist *zsl, double score, robj *obj);
unsigned char *zzlInsert(unsigned char *zl, robj *ele, double score);
int zslDelete(zskiplist *zsl, double score, robj *obj);
zskiplistNode *zslFirstInRange(zskiplist *zsl, zrangespec range);
zskiplistNode *zslFirstInRange(zskiplist *zsl, zrangespec *range);
double zzlGetScore(unsigned char *sptr);
void zzlNext(unsigned char *zl, unsigned char **eptr, unsigned char **sptr);
void zzlPrev(unsigned char *zl, unsigned char **eptr, unsigned char **sptr);
@@ -1102,7 +1114,8 @@ void populateCommandTable(void);
void resetCommandTableStats(void);
void adjustOpenFilesLimit(void);
void closeListeningSockets(int unlink_unix_socket);
void updateCachedTime();
void updateCachedTime(void);
void resetServerStats(void);
/* Set data type */
robj *setTypeCreate(robj *value);
@@ -1174,6 +1187,7 @@ void setKey(redisDb *db, robj *key, robj *val);
int dbExists(redisDb *db, robj *key);
robj *dbRandomKey(redisDb *db);
int dbDelete(redisDb *db, robj *key);
robj *dbUnshareStringValue(redisDb *db, robj *key, robj *o);
long long emptyDb(void(callback)(void*));
int selectDb(redisClient *c, int id);
void signalModifiedKey(redisDb *db, robj *key);
@@ -1292,12 +1306,16 @@ void zincrbyCommand(redisClient *c);
void zrangeCommand(redisClient *c);
void zrangebyscoreCommand(redisClient *c);
void zrevrangebyscoreCommand(redisClient *c);
void zrangebylexCommand(redisClient *c);
void zrevrangebylexCommand(redisClient *c);
void zcountCommand(redisClient *c);
void zlexcountCommand(redisClient *c);
void zrevrangeCommand(redisClient *c);
void zcardCommand(redisClient *c);
void zremCommand(redisClient *c);
void zscoreCommand(redisClient *c);
void zremrangebyscoreCommand(redisClient *c);
void zremrangebylexCommand(redisClient *c);
void multiCommand(redisClient *c);
void execCommand(redisClient *c);
void discardCommand(redisClient *c);
@@ -1348,6 +1366,11 @@ void bitopCommand(redisClient *c);
void bitcountCommand(redisClient *c);
void bitposCommand(redisClient *c);
void replconfCommand(redisClient *c);
void pfselftestCommand(redisClient *c);
void pfaddCommand(redisClient *c);
void pfcountCommand(redisClient *c);
void pfmergeCommand(redisClient *c);
void pfdebugCommand(redisClient *c);
#if defined(__GNUC__)
void *calloc(size_t count, size_t size) __attribute__ ((deprecated));
+113 -26
View File
@@ -200,13 +200,20 @@ void luaSortArray(lua_State *lua) {
lua_pop(lua,1); /* Stack: array (sorted) */
}
#define LUA_CMD_OBJCACHE_SIZE 32
#define LUA_CMD_OBJCACHE_MAX_LEN 64
int luaRedisGenericCommand(lua_State *lua, int raise_error) {
int j, argc = lua_gettop(lua);
struct redisCommand *cmd;
robj **argv;
redisClient *c = server.lua_client;
sds reply;
/* Cached across calls. */
static robj **argv = NULL;
static int argv_size = 0;
static robj *cached_objects[LUA_CMD_OBJCACHE_SIZE];
static int cached_objects_len[LUA_CMD_OBJCACHE_SIZE];
/* Require at least one argument */
if (argc == 0) {
luaPushError(lua,
@@ -215,13 +222,47 @@ int luaRedisGenericCommand(lua_State *lua, int raise_error) {
}
/* Build the arguments vector */
argv = zmalloc(sizeof(robj*)*argc);
for (j = 0; j < argc; j++) {
if (!lua_isstring(lua,j+1)) break;
argv[j] = createStringObject((char*)lua_tostring(lua,j+1),
lua_strlen(lua,j+1));
if (!argv) {
argv = zmalloc(sizeof(robj*)*argc);
} else if (argv_size < argc) {
argv = zrealloc(argv,sizeof(robj*)*argc);
argv_size = argc;
}
for (j = 0; j < argc; j++) {
char *obj_s;
size_t obj_len;
char dbuf[64];
if (lua_isnumber(lua,j+1)) {
/* We can't use lua_tolstring() for number -> string conversion
* since Lua uses a format specifier that loses precision. */
lua_Number num = lua_tonumber(lua,j+1);
obj_len = snprintf(dbuf,sizeof(dbuf),"%.17g",(double)num);
obj_s = dbuf;
} else {
obj_s = (char*)lua_tolstring(lua,j+1,&obj_len);
if (obj_s == NULL) break; /* Not a string. */
}
/* Try to use a cached object. */
if (j < LUA_CMD_OBJCACHE_SIZE && cached_objects[j] &&
cached_objects_len[j] >= obj_len)
{
char *s = cached_objects[j]->ptr;
struct sdshdr *sh = (void*)(s-(sizeof(struct sdshdr)));
argv[j] = cached_objects[j];
cached_objects[j] = NULL;
memcpy(s,obj_s,obj_len+1);
sh->free += sh->len - obj_len;
sh->len = obj_len;
} else {
argv[j] = createStringObject(obj_s, obj_len);
}
}
/* Check if one of the arguments passed by the Lua script
* is not a string or an integer (lua_isstring() return true for
* integers as well). */
@@ -231,7 +272,6 @@ int luaRedisGenericCommand(lua_State *lua, int raise_error) {
decrRefCount(argv[j]);
j--;
}
zfree(argv);
luaPushError(lua,
"Lua redis() command arguments must be strings or integers");
return 1;
@@ -306,16 +346,22 @@ int luaRedisGenericCommand(lua_State *lua, int raise_error) {
/* Convert the result of the Redis command into a suitable Lua type.
* The first thing we need is to create a single string from the client
* output buffers. */
reply = sdsempty();
if (c->bufpos) {
reply = sdscatlen(reply,c->buf,c->bufpos);
if (listLength(c->reply) == 0 && c->bufpos < REDIS_REPLY_CHUNK_BYTES) {
/* This is a fast path for the common case of a reply inside the
* client static buffer. Don't create an SDS string but just use
* the client buffer directly. */
c->buf[c->bufpos] = '\0';
reply = c->buf;
c->bufpos = 0;
}
while(listLength(c->reply)) {
robj *o = listNodeValue(listFirst(c->reply));
} else {
reply = sdsnewlen(c->buf,c->bufpos);
c->bufpos = 0;
while(listLength(c->reply)) {
robj *o = listNodeValue(listFirst(c->reply));
reply = sdscatlen(reply,o->ptr,sdslen(o->ptr));
listDelNode(c->reply,listFirst(c->reply));
reply = sdscatlen(reply,o->ptr,sdslen(o->ptr));
listDelNode(c->reply,listFirst(c->reply));
}
}
if (raise_error && reply[0] != '-') raise_error = 0;
redisProtocolToLuaType(lua,reply);
@@ -325,15 +371,37 @@ int luaRedisGenericCommand(lua_State *lua, int raise_error) {
(reply[0] == '*' && reply[1] != '-')) {
luaSortArray(lua);
}
sdsfree(reply);
if (reply != c->buf) sdsfree(reply);
c->reply_bytes = 0;
cleanup:
/* Clean up. Command code may have changed argv/argc so we use the
* argv/argc of the client instead of the local variables. */
for (j = 0; j < c->argc; j++)
decrRefCount(c->argv[j]);
zfree(c->argv);
for (j = 0; j < c->argc; j++) {
robj *o = c->argv[j];
/* Try to cache the object in the cached_objects array.
* The object must be small, SDS-encoded, and with refcount = 1
* (we must be the only owner) for us to cache it. */
if (j < LUA_CMD_OBJCACHE_SIZE &&
o->refcount == 1 &&
o->encoding == REDIS_ENCODING_RAW &&
sdslen(o->ptr) <= LUA_CMD_OBJCACHE_MAX_LEN)
{
struct sdshdr *sh = (void*)(((char*)(o->ptr))-(sizeof(struct sdshdr)));
if (cached_objects[j]) decrRefCount(cached_objects[j]);
cached_objects[j] = o;
cached_objects_len[j] = sh->free + sh->len;
} else {
decrRefCount(o);
}
}
if (c->argv != argv) {
zfree(c->argv);
argv = NULL;
}
if (raise_error) {
/* If we are here we should have an error in the stack, in the
@@ -452,8 +520,7 @@ void luaMaskCountHook(lua_State *lua, lua_Debug *ar) {
* here when the EVAL command will return. */
aeDeleteFileEvent(server.el, server.lua_caller->fd, AE_READABLE);
}
if (server.lua_timedout)
aeProcessEvents(server.el, AE_FILE_EVENTS|AE_DONT_WAIT);
if (server.lua_timedout) processEventsWhileBlocked();
if (server.lua_kill) {
redisLog(REDIS_WARNING,"Lua script killed by user with SCRIPT KILL.");
lua_pushstring(lua,"Script killed by user with SCRIPT KILL...");
@@ -476,7 +543,7 @@ void luaLoadLibraries(lua_State *lua) {
luaLoadLib(lua, LUA_TABLIBNAME, luaopen_table);
luaLoadLib(lua, LUA_STRLIBNAME, luaopen_string);
luaLoadLib(lua, LUA_MATHLIBNAME, luaopen_math);
luaLoadLib(lua, LUA_DBLIBNAME, luaopen_debug);
luaLoadLib(lua, LUA_DBLIBNAME, luaopen_debug);
luaLoadLib(lua, "cjson", luaopen_cjson);
luaLoadLib(lua, "struct", luaopen_struct);
luaLoadLib(lua, "cmsgpack", luaopen_cmsgpack);
@@ -856,8 +923,12 @@ void evalGenericCommand(redisClient *c, int evalsha) {
int j;
char *sha = c->argv[1]->ptr;
/* Convert to lowercase. We don't use tolower since the function
* managed to always show up in the profiler output consuming
* a non trivial amount of time. */
for (j = 0; j < 40; j++)
funcname[j+2] = tolower(sha[j]);
funcname[j+2] = (sha[j] >= 'A' && sha[j] <= 'Z') ?
sha[j]+('a'-'A') : sha[j];
funcname[42] = '\0';
}
@@ -894,7 +965,7 @@ void evalGenericCommand(redisClient *c, int evalsha) {
/* Select the right DB in the context of the Lua client */
selectDb(server.lua_client,c->db->id);
/* Set a hook in order to be able to stop the script execution if it
* is running for too much time.
* We set the hook only if the time limit is enabled as the hook will
@@ -923,7 +994,23 @@ void evalGenericCommand(redisClient *c, int evalsha) {
}
server.lua_caller = NULL;
selectDb(c,server.lua_client->db->id); /* set DB ID from Lua client */
lua_gc(lua,LUA_GCSTEP,1);
/* Call the Lua garbage collector from time to time to avoid a
* full cycle performed by Lua, which adds too latency.
*
* The call is performed every LUA_GC_CYCLE_PERIOD executed commands
* (and for LUA_GC_CYCLE_PERIOD collection steps) because calling it
* for every command uses too much CPU. */
#define LUA_GC_CYCLE_PERIOD 50
{
static long gc_count = 0;
gc_count++;
if (gc_count == LUA_GC_CYCLE_PERIOD) {
lua_gc(lua,LUA_GCSTEP,LUA_GC_CYCLE_PERIOD);
gc_count = 0;
}
}
if (err) {
addReplyErrorFormat(c,"Error running script (call to %s): %s\n",
+18 -5
View File
@@ -292,24 +292,37 @@ sds sdscpy(sds s, const char *t) {
/* Like sdscatpritf() but gets va_list instead of being variadic. */
sds sdscatvprintf(sds s, const char *fmt, va_list ap) {
va_list cpy;
char *buf, *t;
size_t buflen = 16;
char staticbuf[1024], *buf = staticbuf, *t;
size_t buflen = strlen(fmt)*2;
while(1) {
/* We try to start using a static buffer for speed.
* If not possible we revert to heap allocation. */
if (buflen > sizeof(staticbuf)) {
buf = zmalloc(buflen);
if (buf == NULL) return NULL;
} else {
buflen = sizeof(staticbuf);
}
/* Try with buffers two times bigger every time we fail to
* fit the string in the current buffer size. */
while(1) {
buf[buflen-2] = '\0';
va_copy(cpy,ap);
vsnprintf(buf, buflen, fmt, cpy);
if (buf[buflen-2] != '\0') {
zfree(buf);
if (buf != staticbuf) zfree(buf);
buflen *= 2;
buf = zmalloc(buflen);
if (buf == NULL) return NULL;
continue;
}
break;
}
/* Finally concat the obtained string to the SDS string and return it. */
t = sdscat(s, buf);
zfree(buf);
if (buf != staticbuf) zfree(buf);
return t;
}
+296 -139
View File
@@ -88,7 +88,7 @@ typedef struct sentinelAddr {
/* Failover machine different states. */
#define SENTINEL_FAILOVER_STATE_NONE 0 /* No failover in progress. */
#define SENTINEL_FAILOVER_STATE_WAIT_START 1 /* Wait for failover_start_time*/
#define SENTINEL_FAILOVER_STATE_WAIT_START 1 /* Wait for failover_start_time*/
#define SENTINEL_FAILOVER_STATE_SELECT_SLAVE 2 /* Select slave to promote */
#define SENTINEL_FAILOVER_STATE_SEND_SLAVEOF_NOONE 3 /* Slave -> Master */
#define SENTINEL_FAILOVER_STATE_WAIT_PROMOTION 4 /* Wait slave to change role */
@@ -129,6 +129,10 @@ typedef struct sentinelRedisInstance {
mstime_t pc_last_activity; /* Last time we received any message. */
mstime_t last_avail_time; /* Last time the instance replied to ping with
a reply we consider valid. */
mstime_t last_ping_time; /* Last time a pending ping was sent in the
context of the current command connection
with the instance. 0 if still not sent or
if pong already received. */
mstime_t last_pong_time; /* Last time the instance replied to ping,
whatever the reply was. That's used to check
if the link is idle and must be reconnected. */
@@ -179,6 +183,8 @@ typedef struct sentinelRedisInstance {
mstime_t failover_state_change_time;
mstime_t failover_start_time; /* Last failover attempt start time. */
mstime_t failover_timeout; /* Max time to refresh failover state. */
mstime_t failover_delay_logged; /* For what failover_start_time value we
logged the failover delay. */
struct sentinelRedisInstance *promoted_slave; /* Promoted slave instance. */
/* Scripts executed to notify admin or reconfigure clients: when they
* are set to NULL no script is executed. */
@@ -329,6 +335,7 @@ int sentinelSendSlaveOf(sentinelRedisInstance *ri, char *host, int port);
char *sentinelVoteLeader(sentinelRedisInstance *master, uint64_t req_epoch, char *req_runid, uint64_t *leader_epoch);
void sentinelFlushConfig(void);
void sentinelGenerateInitialMonitorEvents(void);
int sentinelSendPing(sentinelRedisInstance *ri);
/* ========================= Dictionary types =============================== */
@@ -371,6 +378,7 @@ dictType leaderVotesDictType = {
void sentinelCommand(redisClient *c);
void sentinelInfoCommand(redisClient *c);
void sentinelSetCommand(redisClient *c);
void sentinelPublishCommand(redisClient *c);
struct redisCommand sentinelcmds[] = {
{"ping",pingCommand,1,"",0,NULL,0,0,0,0,0},
@@ -379,6 +387,7 @@ struct redisCommand sentinelcmds[] = {
{"unsubscribe",unsubscribeCommand,-1,"",0,NULL,0,0,0,0,0},
{"psubscribe",psubscribeCommand,-2,"",0,NULL,0,0,0,0,0},
{"punsubscribe",punsubscribeCommand,-1,"",0,NULL,0,0,0,0,0},
{"publish",sentinelPublishCommand,3,"",0,NULL,0,0,0,0,0},
{"info",sentinelInfoCommand,-1,"",0,NULL,0,0,0,0,0},
{"shutdown",shutdownCommand,-1,"",0,NULL,0,0,0,0,0}
};
@@ -484,7 +493,7 @@ int sentinelAddrIsEqual(sentinelAddr *a, sentinelAddr *b) {
/* =========================== Events notification ========================== */
/* Send an event to log, pub/sub, user notification script.
*
*
* 'level' is the log level for logging. Only REDIS_WARNING events will trigger
* the execution of the user notification script.
*
@@ -607,7 +616,7 @@ void sentinelScheduleScriptExecution(char *path, ...) {
}
va_end(ap);
argv[0] = sdsnew(path);
sj = zmalloc(sizeof(*sj));
sj->flags = SENTINEL_SCRIPT_NONE;
sj->retry_num = 0;
@@ -733,7 +742,7 @@ void sentinelCollectTerminatedScripts(void) {
if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);
sentinelEvent(REDIS_DEBUG,"-script-child",NULL,"%ld %d %d",
(long)pid, exitcode, bysignal);
ln = sentinelGetScriptListNodeByPid(pid);
if (ln == NULL) {
redisLog(REDIS_WARNING,"wait3() returned a pid (%ld) we can't find in our scripts execution queue!", (long)pid);
@@ -923,6 +932,11 @@ sentinelRedisInstance *createSentinelRedisInstance(char *name, int flags, char *
ri->cc_conn_time = 0;
ri->pc_conn_time = 0;
ri->pc_last_activity = 0;
/* We set the last_ping_time to "now" even if we actually don't have yet
* a connection with the node, nor we sent a ping.
* This is useful to detect a timeout in case we'll not be able to connect
* with the node at all. */
ri->last_ping_time = mstime();
ri->last_avail_time = mstime();
ri->last_pong_time = mstime();
ri->last_pub_time = mstime();
@@ -955,6 +969,7 @@ sentinelRedisInstance *createSentinelRedisInstance(char *name, int flags, char *
ri->failover_state_change_time = 0;
ri->failover_start_time = 0;
ri->failover_timeout = SENTINEL_DEFAULT_FAILOVER_TIMEOUT;
ri->failover_delay_logged = 0;
ri->promoted_slave = NULL;
ri->notification_script = NULL;
ri->client_reconfig_script = NULL;
@@ -1005,7 +1020,7 @@ sentinelRedisInstance *sentinelRedisInstanceLookupSlave(
{
sds key;
sentinelRedisInstance *slave;
redisAssert(ri->flags & SRI_MASTER);
key = sdscatprintf(sdsempty(),
strchr(ip,':') ? "[%s]:%d" : "%s:%d",
@@ -1025,7 +1040,7 @@ const char *sentinelRedisInstanceTypeStr(sentinelRedisInstance *ri) {
/* This function removes all the instances found in the dictionary of
* sentinels in the specified 'master', having either:
*
*
* 1) The same ip/port as specified.
* 2) The same runid.
*
@@ -1159,6 +1174,7 @@ void sentinelResetMaster(sentinelRedisInstance *ri, int flags) {
sdsfree(ri->slave_master_host);
ri->runid = NULL;
ri->slave_master_host = NULL;
ri->last_ping_time = mstime();
ri->last_avail_time = mstime();
ri->last_pong_time = mstime();
ri->role_reported_time = mstime();
@@ -1218,7 +1234,7 @@ int sentinelResetMasterAndChangeAddress(sentinelRedisInstance *master, char *ip,
slave->addr->port);
}
dictReleaseIterator(di);
/* If we are switching to a different address, include the old address
* as a slave as well, so that we'll be able to sense / reconfigure
* the old master. */
@@ -1285,6 +1301,24 @@ sentinelAddr *sentinelGetCurrentMasterAddress(sentinelRedisInstance *master) {
}
}
/* This function sets the down_after_period field value in 'master' to all
* the slaves and sentinel instances connected to this master. */
void sentinelPropagateDownAfterPeriod(sentinelRedisInstance *master) {
dictIterator *di;
dictEntry *de;
int j;
dict *d[] = {master->slaves, master->sentinels, NULL};
for (j = 0; d[j]; j++) {
di = dictGetIterator(d[j]);
while((de = dictNext(di)) != NULL) {
sentinelRedisInstance *ri = dictGetVal(de);
ri->down_after_period = master->down_after_period;
}
dictReleaseIterator(di);
}
}
/* ============================ Config handling ============================= */
char *sentinelHandleConfiguration(char **argv, int argc) {
sentinelRedisInstance *ri;
@@ -1310,6 +1344,7 @@ char *sentinelHandleConfiguration(char **argv, int argc) {
ri->down_after_period = atoi(argv[2]);
if (ri->down_after_period <= 0)
return "negative or zero time parameter.";
sentinelPropagateDownAfterPeriod(ri);
} else if (!strcasecmp(argv[0],"failover-timeout") && argc == 3) {
/* failover-timeout <name> <milliseconds> */
ri = sentinelGetMasterByName(argv[1]);
@@ -1342,13 +1377,26 @@ char *sentinelHandleConfiguration(char **argv, int argc) {
ri = sentinelGetMasterByName(argv[1]);
if (!ri) return "No such master with specified name.";
ri->auth_pass = sdsnew(argv[2]);
} else if (!strcasecmp(argv[0],"current-epoch") && argc == 2) {
/* current-epoch <epoch> */
unsigned long long current_epoch = strtoull(argv[1],NULL,10);
if (current_epoch > sentinel.current_epoch)
sentinel.current_epoch = current_epoch;
} else if (!strcasecmp(argv[0],"config-epoch") && argc == 3) {
/* config-epoch <name> <epoch> */
ri = sentinelGetMasterByName(argv[1]);
if (!ri) return "No such master with specified name.";
ri->config_epoch = strtoull(argv[2],NULL,10);
/* The following update of current_epoch is not really useful as
* now the current epoch is persisted on the config file, but
* we leave this check here for redundancy. */
if (ri->config_epoch > sentinel.current_epoch)
sentinel.current_epoch = ri->config_epoch;
} else if (!strcasecmp(argv[0],"leader-epoch") && argc == 3) {
/* leader-epoch <name> <epoch> */
ri = sentinelGetMasterByName(argv[1]);
if (!ri) return "No such master with specified name.";
ri->leader_epoch = strtoull(argv[2],NULL,10);
} else if (!strcasecmp(argv[0],"known-slave") && argc == 4) {
sentinelRedisInstance *slave;
@@ -1387,13 +1435,13 @@ char *sentinelHandleConfiguration(char **argv, int argc) {
void rewriteConfigSentinelOption(struct rewriteConfigState *state) {
dictIterator *di, *di2;
dictEntry *de;
sds line;
/* For every master emit a "sentinel monitor" config entry. */
di = dictGetIterator(sentinel.masters);
while((de = dictNext(di)) != NULL) {
sentinelRedisInstance *master, *ri;
sentinelAddr *master_addr;
sds line;
/* sentinel monitor */
master = dictGetVal(de);
@@ -1457,6 +1505,12 @@ void rewriteConfigSentinelOption(struct rewriteConfigState *state) {
master->name, (unsigned long long) master->config_epoch);
rewriteConfigRewriteLine(state,"sentinel",line,1);
/* sentinel leader-epoch */
line = sdscatprintf(sdsempty(),
"sentinel leader-epoch %s %llu",
master->name, (unsigned long long) master->leader_epoch);
rewriteConfigRewriteLine(state,"sentinel",line,1);
/* sentinel known-slave */
di2 = dictGetIterator(master->slaves);
while((de = dictNext(di2)) != NULL) {
@@ -1492,6 +1546,12 @@ void rewriteConfigSentinelOption(struct rewriteConfigState *state) {
}
dictReleaseIterator(di2);
}
/* sentinel current-epoch is a global state valid for all the masters. */
line = sdscatprintf(sdsempty(),
"sentinel current-epoch %llu", (unsigned long long) sentinel.current_epoch);
rewriteConfigRewriteLine(state,"sentinel",line,1);
dictReleaseIterator(di);
}
@@ -1503,21 +1563,23 @@ void rewriteConfigSentinelOption(struct rewriteConfigState *state) {
*
* On failure the function logs a warning on the Redis log. */
void sentinelFlushConfig(void) {
int fd;
int fd = -1;
int saved_hz = server.hz;
int rewrite_status;
server.hz = REDIS_DEFAULT_HZ;
if (rewriteConfig(server.configfile) != -1) {
/* Rewrite succeded, fsync it. */
if ((fd = open(server.configfile,O_RDONLY)) != -1) {
fsync(fd);
close(fd);
}
} else {
redisLog(REDIS_WARNING,"WARNING: Senitnel was not able to save the new configuration on disk!!!: %s", strerror(errno));
}
rewrite_status = rewriteConfig(server.configfile);
server.hz = saved_hz;
if (rewrite_status == -1) goto werr;
if ((fd = open(server.configfile,O_RDONLY)) == -1) goto werr;
if (fsync(fd) == -1) goto werr;
if (close(fd) == EOF) goto werr;
return;
werr:
if (fd != -1) close(fd);
redisLog(REDIS_WARNING,"WARNING: Sentinel was not able to save the new configuration on disk!!!: %s", strerror(errno));
}
/* ====================== hiredis connection handling ======================= */
@@ -1588,6 +1650,23 @@ void sentinelSendAuthIfNeeded(sentinelRedisInstance *ri, redisAsyncContext *c) {
}
}
/* Use CLIENT SETNAME to name the connection in the Redis instance as
* sentinel-<first_8_chars_of_runid>-<connection_type>
* The connection type is "cmd" or "pubsub" as specified by 'type'.
*
* This makes it possible to list all the sentinel instances connected
* to a Redis servewr with CLIENT LIST, grepping for a specific name format. */
void sentinelSetClientName(sentinelRedisInstance *ri, redisAsyncContext *c, char *type) {
char name[64];
snprintf(name,sizeof(name),"sentinel-%.8s-%s",server.runid,type);
if (redisAsyncCommand(c, sentinelDiscardReplyCallback, NULL,
"CLIENT SETNAME %s", name) == REDIS_OK)
{
ri->pending_commands++;
}
}
/* Create the async connections for the specified instance if the instance
* is disconnected. Note that the SRI_DISCONNECTED flag is set even if just
* one of the two links (commands and pub/sub) is missing. */
@@ -1610,6 +1689,10 @@ void sentinelReconnectInstance(sentinelRedisInstance *ri) {
redisAsyncSetDisconnectCallback(ri->cc,
sentinelDisconnectCallback);
sentinelSendAuthIfNeeded(ri,ri->cc);
sentinelSetClientName(ri,ri->cc,"cmd");
/* Send a PING ASAP when reconnecting. */
sentinelSendPing(ri);
}
}
/* Pub / Sub */
@@ -1630,6 +1713,7 @@ void sentinelReconnectInstance(sentinelRedisInstance *ri) {
redisAsyncSetDisconnectCallback(ri->pc,
sentinelDisconnectCallback);
sentinelSendAuthIfNeeded(ri,ri->pc);
sentinelSetClientName(ri,ri->pc,"pubsub");
/* Now we subscribe to the Sentinels "Hello" channel. */
retval = redisAsyncCommand(ri->pc,
sentinelReceiveHelloMessages, NULL, "SUBSCRIBE %s",
@@ -1763,7 +1847,7 @@ void sentinelRefreshInstanceInfo(sentinelRedisInstance *ri, const char *info) {
ri->slave_conf_change_time = mstime();
}
}
/* master_link_status:<status> */
if (sdslen(l) >= 19 && !memcmp(l,"master_link_status:",19)) {
ri->slave_master_link_status =
@@ -1803,6 +1887,10 @@ void sentinelRefreshInstanceInfo(sentinelRedisInstance *ri, const char *info) {
ri->flags & SRI_MASTER ? "master" : "slave");
}
/* None of the following conditions are processed when in tilt mode, so
* return asap. */
if (sentinel.tilt) return;
/* Handle master -> slave role switch. */
if ((ri->flags & SRI_MASTER) && role == SRI_SLAVE) {
/* Nothing to do, but masters claiming to be slaves are
@@ -1814,8 +1902,7 @@ void sentinelRefreshInstanceInfo(sentinelRedisInstance *ri, const char *info) {
if ((ri->flags & SRI_SLAVE) && role == SRI_MASTER) {
/* If this is a promoted slave we can change state to the
* failover state machine. */
if (!sentinel.tilt &&
(ri->master->flags & SRI_FAILOVER_IN_PROGRESS) &&
if ((ri->master->flags & SRI_FAILOVER_IN_PROGRESS) &&
(ri->master->failover_state ==
SENTINEL_FAILOVER_STATE_WAIT_PROMOTION))
{
@@ -1833,7 +1920,7 @@ void sentinelRefreshInstanceInfo(sentinelRedisInstance *ri, const char *info) {
ri->master,"%@");
sentinelCallClientReconfScript(ri->master,SENTINEL_LEADER,
"start",ri->master->addr,ri->addr);
} else if (!sentinel.tilt) {
} else {
/* A slave turned into a master. We want to force our view and
* reconfigure as slave. Wait some time after the change before
* going forward, to receive new configs if any. */
@@ -1853,7 +1940,7 @@ void sentinelRefreshInstanceInfo(sentinelRedisInstance *ri, const char *info) {
}
/* Handle slaves replicating to a different master address. */
if ((ri->flags & SRI_SLAVE) && !sentinel.tilt &&
if ((ri->flags & SRI_SLAVE) &&
role == SRI_SLAVE &&
(ri->slave_master_port != ri->master->addr->port ||
strcasecmp(ri->slave_master_host,ri->master->addr->ip)))
@@ -1874,10 +1961,6 @@ void sentinelRefreshInstanceInfo(sentinelRedisInstance *ri, const char *info) {
}
}
/* None of the following conditions are processed when in tilt mode, so
* return asap. */
if (sentinel.tilt) return;
/* Detect if the slave that is in the process of being reconfigured
* changed state. */
if ((ri->flags & SRI_SLAVE) && role == SRI_SLAVE &&
@@ -1944,6 +2027,7 @@ void sentinelPingReplyCallback(redisAsyncContext *c, void *reply, void *privdata
strncmp(r->str,"MASTERDOWN",10) == 0)
{
ri->last_avail_time = mstime();
ri->last_ping_time = 0; /* Flag the pong as received. */
} else {
/* Send a SCRIPT KILL command if the instance appears to be
* down because of a busy script. */
@@ -1978,22 +2062,113 @@ void sentinelPublishReplyCallback(redisAsyncContext *c, void *reply, void *privd
ri->last_pub_time = mstime();
}
/* Process an hello message received via Pub/Sub in master or slave instance,
* or sent directly to this sentinel via the (fake) PUBLISH command of Sentinel.
*
* If the master name specified in the message is not known, the message is
* discareded. */
void sentinelProcessHelloMessage(char *hello, int hello_len) {
/* Format is composed of 8 tokens:
* 0=ip,1=port,2=runid,3=current_epoch,4=master_name,
* 5=master_ip,6=master_port,7=master_config_epoch. */
int numtokens, port, removed, master_port;
uint64_t current_epoch, master_config_epoch;
char **token = sdssplitlen(hello, hello_len, ",", 1, &numtokens);
sentinelRedisInstance *si, *master;
if (numtokens == 8) {
/* Obtain a reference to the master this hello message is about */
master = sentinelGetMasterByName(token[4]);
if (!master) goto cleanup; /* Unknown master, skip the message. */
/* First, try to see if we already have this sentinel. */
port = atoi(token[1]);
master_port = atoi(token[6]);
si = getSentinelRedisInstanceByAddrAndRunID(
master->sentinels,token[0],port,token[2]);
current_epoch = strtoull(token[3],NULL,10);
master_config_epoch = strtoull(token[7],NULL,10);
if (!si) {
/* If not, remove all the sentinels that have the same runid
* OR the same ip/port, because it's either a restart or a
* network topology change. */
removed = removeMatchingSentinelsFromMaster(master,token[0],port,
token[2]);
if (removed) {
sentinelEvent(REDIS_NOTICE,"-dup-sentinel",master,
"%@ #duplicate of %s:%d or %s",
token[0],port,token[2]);
}
/* Add the new sentinel. */
si = createSentinelRedisInstance(NULL,SRI_SENTINEL,
token[0],port,master->quorum,master);
if (si) {
sentinelEvent(REDIS_NOTICE,"+sentinel",si,"%@");
/* The runid is NULL after a new instance creation and
* for Sentinels we don't have a later chance to fill it,
* so do it now. */
si->runid = sdsnew(token[2]);
sentinelFlushConfig();
}
}
/* Update local current_epoch if received current_epoch is greater.*/
if (current_epoch > sentinel.current_epoch) {
sentinel.current_epoch = current_epoch;
sentinelFlushConfig();
sentinelEvent(REDIS_WARNING,"+new-epoch",master,"%llu",
(unsigned long long) sentinel.current_epoch);
}
/* Update master info if received configuration is newer. */
if (master->config_epoch < master_config_epoch) {
master->config_epoch = master_config_epoch;
if (master_port != master->addr->port ||
strcmp(master->addr->ip, token[5]))
{
sentinelAddr *old_addr;
sentinelEvent(REDIS_WARNING,"+config-update-from",si,"%@");
sentinelEvent(REDIS_WARNING,"+switch-master",
master,"%s %s %d %s %d",
master->name,
master->addr->ip, master->addr->port,
token[5], master_port);
old_addr = dupSentinelAddr(master->addr);
sentinelResetMasterAndChangeAddress(master, token[5], master_port);
sentinelCallClientReconfScript(master,
SENTINEL_OBSERVER,"start",
old_addr,master->addr);
releaseSentinelAddr(old_addr);
}
}
/* Update the state of the Sentinel. */
if (si) si->last_hello_time = mstime();
}
cleanup:
sdsfreesplitres(token,numtokens);
}
/* This is our Pub/Sub callback for the Hello channel. It's useful in order
* to discover other sentinels attached at the same master. */
void sentinelReceiveHelloMessages(redisAsyncContext *c, void *reply, void *privdata) {
sentinelRedisInstance *ri = c->data, *master;
sentinelRedisInstance *ri = c->data;
redisReply *r;
if (!reply || !ri) return;
r = reply;
master = (ri->flags & SRI_MASTER) ? ri : ri->master;
/* Update the last activity in the pubsub channel. Note that since we
* receive our messages as well this timestamp can be used to detect
* if the link is probably disconnected even if it seems otherwise. */
ri->pc_last_activity = mstime();
/* Sanity check in the reply we expect, so that the code that follows
* can avoid to check for details. */
if (r->type != REDIS_REPLY_ARRAY ||
@@ -2006,90 +2181,7 @@ void sentinelReceiveHelloMessages(redisAsyncContext *c, void *reply, void *privd
/* We are not interested in meeting ourselves */
if (strstr(r->element[2]->str,server.runid) != NULL) return;
{
/* Format is composed of 8 tokens:
* 0=ip,1=port,2=runid,3=current_epoch,4=master_name,
* 5=master_ip,6=master_port,7=master_config_epoch. */
int numtokens, port, removed, master_port;
uint64_t current_epoch, master_config_epoch;
char **token = sdssplitlen(r->element[2]->str,
r->element[2]->len,
",",1,&numtokens);
sentinelRedisInstance *si;
if (numtokens == 8) {
/* First, try to see if we already have this sentinel. */
port = atoi(token[1]);
master_port = atoi(token[6]);
si = getSentinelRedisInstanceByAddrAndRunID(
master->sentinels,token[0],port,token[2]);
current_epoch = strtoull(token[3],NULL,10);
master_config_epoch = strtoull(token[7],NULL,10);
sentinelRedisInstance *msgmaster;
if (!si) {
/* If not, remove all the sentinels that have the same runid
* OR the same ip/port, because it's either a restart or a
* network topology change. */
removed = removeMatchingSentinelsFromMaster(master,token[0],port,
token[2]);
if (removed) {
sentinelEvent(REDIS_NOTICE,"-dup-sentinel",master,
"%@ #duplicate of %s:%d or %s",
token[0],port,token[2]);
}
/* Add the new sentinel. */
si = createSentinelRedisInstance(NULL,SRI_SENTINEL,
token[0],port,master->quorum,master);
if (si) {
sentinelEvent(REDIS_NOTICE,"+sentinel",si,"%@");
/* The runid is NULL after a new instance creation and
* for Sentinels we don't have a later chance to fill it,
* so do it now. */
si->runid = sdsnew(token[2]);
sentinelFlushConfig();
}
}
/* Update local current_epoch if received current_epoch is greater.*/
if (current_epoch > sentinel.current_epoch) {
sentinel.current_epoch = current_epoch;
sentinelEvent(REDIS_WARNING,"+new-epoch",ri,"%llu",
(unsigned long long) sentinel.current_epoch);
}
/* Update master info if received configuration is newer. */
if ((msgmaster = sentinelGetMasterByName(token[4])) != NULL) {
if (msgmaster->config_epoch < master_config_epoch) {
msgmaster->config_epoch = master_config_epoch;
if (master_port != msgmaster->addr->port ||
strcmp(msgmaster->addr->ip, token[5]))
{
sentinelAddr *old_addr;
sentinelEvent(REDIS_WARNING,"+switch-master",
msgmaster,"%s %s %d %s %d",
msgmaster->name,
msgmaster->addr->ip, msgmaster->addr->port,
token[5], master_port);
old_addr = dupSentinelAddr(msgmaster->addr);
sentinelResetMasterAndChangeAddress(msgmaster,
token[5], master_port);
sentinelCallClientReconfScript(msgmaster,
SENTINEL_OBSERVER,"start",
old_addr,msgmaster->addr);
releaseSentinelAddr(old_addr);
}
}
}
/* Update the state of the Sentinel. */
if (si) si->last_hello_time = mstime();
}
sdsfreesplitres(token,numtokens);
}
sentinelProcessHelloMessage(r->element[2]->str, r->element[2]->len);
}
/* Send an "Hello" message via Pub/Sub to the specified 'ri' Redis
@@ -2131,11 +2223,31 @@ int sentinelSendHello(sentinelRedisInstance *ri) {
return REDIS_OK;
}
/* Send a PING to the specified instance and refresh the last_ping_time
* if it is zero (that is, if we received a pong for the previous ping).
*
* On error zero is returned, and we can't consider the PING command
* queued in the connection. */
int sentinelSendPing(sentinelRedisInstance *ri) {
int retval = redisAsyncCommand(ri->cc,
sentinelPingReplyCallback, NULL, "PING");
if (retval == REDIS_OK) {
ri->pending_commands++;
/* We update the ping time only if we received the pong for
* the previous ping, otherwise we are technically waiting
* since the first ping that did not received a reply. */
if (ri->last_ping_time == 0) ri->last_ping_time = mstime();
return 1;
} else {
return 0;
}
}
/* Send periodic PING, INFO, and PUBLISH to the Hello channel to
* the specified master or slave instance. */
void sentinelPingInstance(sentinelRedisInstance *ri) {
void sentinelSendPeriodicCommands(sentinelRedisInstance *ri) {
mstime_t now = mstime();
mstime_t info_period;
mstime_t info_period, ping_period;
int retval;
/* Return ASAP if we have already a PING or INFO already pending, or
@@ -2161,6 +2273,12 @@ void sentinelPingInstance(sentinelRedisInstance *ri) {
info_period = SENTINEL_INFO_PERIOD;
}
/* We ping instances every time the last received pong is older than
* the configured 'down-after-milliseconds' time, but every second
* anyway if 'down-after-milliseconds' is greater than 1 second. */
ping_period = ri->down_after_period;
if (ping_period > SENTINEL_PING_PERIOD) ping_period = SENTINEL_PING_PERIOD;
if ((ri->flags & SRI_SENTINEL) == 0 &&
(ri->info_refresh == 0 ||
(now - ri->info_refresh) > info_period))
@@ -2168,18 +2286,12 @@ void sentinelPingInstance(sentinelRedisInstance *ri) {
/* Send INFO to masters and slaves, not sentinels. */
retval = redisAsyncCommand(ri->cc,
sentinelInfoReplyCallback, NULL, "INFO");
if (retval != REDIS_OK) return;
ri->pending_commands++;
} else if ((now - ri->last_pong_time) > SENTINEL_PING_PERIOD) {
if (retval == REDIS_OK) ri->pending_commands++;
} else if ((now - ri->last_pong_time) > ping_period) {
/* Send PING to all the three kinds of instances. */
retval = redisAsyncCommand(ri->cc,
sentinelPingReplyCallback, NULL, "PING");
if (retval != REDIS_OK) return;
ri->pending_commands++;
} else if ((ri->flags & SRI_SENTINEL) == 0 &&
(now - ri->last_pub_time) > SENTINEL_PUBLISH_PERIOD)
{
/* PUBLISH hello messages to masters and slaves. */
sentinelSendPing(ri);
} else if ((now - ri->last_pub_time) > SENTINEL_PUBLISH_PERIOD) {
/* PUBLISH hello messages to all the three kinds of instances. */
sentinelSendHello(ri);
}
}
@@ -2253,6 +2365,11 @@ void addReplySentinelRedisInstance(redisClient *c, sentinelRedisInstance *ri) {
fields++;
}
addReplyBulkCString(c,"last-ping-sent");
addReplyBulkLongLong(c,
ri->last_ping_time ? (mstime() - ri->last_ping_time) : 0);
fields++;
addReplyBulkCString(c,"last-ok-ping-reply");
addReplyBulkLongLong(c,mstime() - ri->last_avail_time);
fields++;
@@ -2273,6 +2390,10 @@ void addReplySentinelRedisInstance(redisClient *c, sentinelRedisInstance *ri) {
fields++;
}
addReplyBulkCString(c,"down-after-milliseconds");
addReplyBulkLongLong(c,ri->down_after_period);
fields++;
/* Masters and Slaves */
if (ri->flags & (SRI_MASTER|SRI_SLAVE)) {
addReplyBulkCString(c,"info-refresh");
@@ -2307,10 +2428,6 @@ void addReplySentinelRedisInstance(redisClient *c, sentinelRedisInstance *ri) {
addReplyBulkLongLong(c,ri->quorum);
fields++;
addReplyBulkCString(c,"down-after-milliseconds");
addReplyBulkLongLong(c,ri->down_after_period);
fields++;
addReplyBulkCString(c,"failover-timeout");
addReplyBulkLongLong(c,ri->failover_timeout);
fields++;
@@ -2667,6 +2784,7 @@ void sentinelSetCommand(redisClient *c) {
if (getLongLongFromObject(o,&ll) == REDIS_ERR || ll <= 0)
goto badfmt;
ri->down_after_period = ll;
sentinelPropagateDownAfterPeriod(ri);
changes++;
} else if (!strcasecmp(option,"failover-timeout")) {
/* failover-timeout <milliseconds> */
@@ -2733,20 +2851,42 @@ badfmt: /* Bad format errors */
value, option);
}
/* Our fake PUBLISH command: it is actually useful only to receive hello messages
* from the other sentinel instances, and publishing to a channel other than
* SENTINEL_HELLO_CHANNEL is forbidden.
*
* Because we have a Sentinel PUBLISH, the code to send hello messages is the same
* for all the three kind of instances: masters, slaves, sentinels. */
void sentinelPublishCommand(redisClient *c) {
if (strcmp(c->argv[1]->ptr,SENTINEL_HELLO_CHANNEL)) {
addReplyError(c, "Only HELLO messages are accepted by Sentinel instances.");
return;
}
sentinelProcessHelloMessage(c->argv[2]->ptr,sdslen(c->argv[2]->ptr));
addReplyLongLong(c,1);
}
/* ===================== SENTINEL availability checks ======================= */
/* Is this instance down from our point of view? */
void sentinelCheckSubjectivelyDown(sentinelRedisInstance *ri) {
mstime_t elapsed = mstime() - ri->last_avail_time;
mstime_t elapsed = 0;
/* Check if we are in need for a reconnection of one of the
if (ri->last_ping_time)
elapsed = mstime() - ri->last_ping_time;
/* Check if we are in need for a reconnection of one of the
* links, because we are detecting low activity.
*
* 1) Check if the command link seems connected, was connected not less
* than SENTINEL_MIN_LINK_RECONNECT_PERIOD, but still we have an
* idle time that is greater than down_after_period / 2 seconds. */
* than SENTINEL_MIN_LINK_RECONNECT_PERIOD, but still we have a
* pending ping for more than half the timeout. */
if (ri->cc &&
(mstime() - ri->cc_conn_time) > SENTINEL_MIN_LINK_RECONNECT_PERIOD &&
ri->last_ping_time != 0 && /* Ther is a pending ping... */
/* The pending ping is delayed, and we did not received
* error replies as well. */
(mstime() - ri->last_ping_time) > (ri->down_after_period/2) &&
(mstime() - ri->last_pong_time) > (ri->down_after_period/2))
{
sentinelKillLink(ri,ri->cc);
@@ -2929,6 +3069,7 @@ void sentinelAskMasterStateToOtherSentinels(sentinelRedisInstance *master, int f
char *sentinelVoteLeader(sentinelRedisInstance *master, uint64_t req_epoch, char *req_runid, uint64_t *leader_epoch) {
if (req_epoch > sentinel.current_epoch) {
sentinel.current_epoch = req_epoch;
sentinelFlushConfig();
sentinelEvent(REDIS_WARNING,"+new-epoch",master,"%llu",
(unsigned long long) sentinel.current_epoch);
}
@@ -2938,6 +3079,7 @@ char *sentinelVoteLeader(sentinelRedisInstance *master, uint64_t req_epoch, char
sdsfree(master->leader);
master->leader = sdsnew(req_runid);
master->leader_epoch = sentinel.current_epoch;
sentinelFlushConfig();
sentinelEvent(REDIS_WARNING,"+vote-for-leader",master,"%s %llu",
master->leader, (unsigned long long) master->leader_epoch);
/* If we did not voted for ourselves, set the master failover start
@@ -3099,7 +3241,7 @@ void sentinelStartFailover(sentinelRedisInstance *master) {
* 1) Master must be in ODOWN condition.
* 2) No failover already in progress.
* 3) No failover already attempted recently.
*
*
* We still don't know if we'll win the election so it is possible that we
* start the failover but that we'll not be able to act.
*
@@ -3113,7 +3255,22 @@ int sentinelStartFailoverIfNeeded(sentinelRedisInstance *master) {
/* Last failover attempt started too little time ago? */
if (mstime() - master->failover_start_time <
master->failover_timeout*2) return 0;
master->failover_timeout*2)
{
if (master->failover_delay_logged != master->failover_start_time) {
time_t clock = (master->failover_start_time +
master->failover_timeout*2) / 1000;
char ctimebuf[26];
ctime_r(&clock,ctimebuf);
ctimebuf[24] = '\0'; /* Remove newline. */
master->failover_delay_logged = master->failover_start_time;
redisLog(REDIS_WARNING,
"Next failover delay: I will not start a failover before %s",
ctimebuf);
}
return 0;
}
sentinelStartFailover(master);
return 1;
@@ -3500,7 +3657,7 @@ void sentinelHandleRedisInstance(sentinelRedisInstance *ri) {
/* ========== MONITORING HALF ============ */
/* Every kind of instance */
sentinelReconnectInstance(ri);
sentinelPingInstance(ri);
sentinelSendPeriodicCommands(ri);
/* ============== ACTING HALF ============= */
/* We don't proceed with the acting half if we are in TILT mode.
+9 -2
View File
@@ -231,8 +231,15 @@ void sortCommand(redisClient *c) {
} else if (!strcasecmp(c->argv[j]->ptr,"alpha")) {
alpha = 1;
} else if (!strcasecmp(c->argv[j]->ptr,"limit") && leftargs >= 2) {
if ((getLongFromObjectOrReply(c, c->argv[j+1], &limit_start, NULL) != REDIS_OK) ||
(getLongFromObjectOrReply(c, c->argv[j+2], &limit_count, NULL) != REDIS_OK)) return;
if ((getLongFromObjectOrReply(c, c->argv[j+1], &limit_start, NULL)
!= REDIS_OK) ||
(getLongFromObjectOrReply(c, c->argv[j+2], &limit_count, NULL)
!= REDIS_OK))
{
decrRefCount(sortval);
listRelease(operations);
return;
}
j+=2;
} else if (!strcasecmp(c->argv[j]->ptr,"store") && leftargs >= 1) {
storekey = c->argv[j+1];
+2 -14
View File
@@ -215,12 +215,7 @@ void setrangeCommand(redisClient *c) {
return;
/* Create a copy when the object is shared or encoded. */
if (o->refcount != 1 || o->encoding != REDIS_ENCODING_RAW) {
robj *decoded = getDecodedObject(o);
o = createStringObject(decoded->ptr, sdslen(decoded->ptr));
decrRefCount(decoded);
dbOverwrite(c->db,c->argv[1],o);
}
o = dbUnshareStringValue(c->db,c->argv[1],o);
}
if (sdslen(value) > 0) {
@@ -433,15 +428,8 @@ void appendCommand(redisClient *c) {
if (checkStringLength(c,totlen) != REDIS_OK)
return;
/* If the object is shared or encoded, we have to make a copy */
if (o->refcount != 1 || o->encoding != REDIS_ENCODING_RAW) {
robj *decoded = getDecodedObject(o);
o = createStringObject(decoded->ptr, sdslen(decoded->ptr));
decrRefCount(decoded);
dbOverwrite(c->db,c->argv[1],o);
}
/* Append the value */
o = dbUnshareStringValue(c->db,c->argv[1],o);
o->ptr = sdscatlen(o->ptr,append->ptr,sdslen(append->ptr));
totlen = sdslen(o->ptr);
}
+706 -95
View File
File diff suppressed because it is too large Load Diff
+1 -1
View File
@@ -402,7 +402,7 @@ void getRandomHexChars(char *p, unsigned int len) {
/* Turn it into hex digits taking just 4 bits out of 8 for every byte. */
for (j = 0; j < len; j++)
p[j] = charset[p[j] & 0x0F];
fclose(fp);
if (fp) fclose(fp);
}
/* Given the filename, return the absolute path as an SDS string, or NULL
+1 -1
View File
@@ -1 +1 @@
#define REDIS_VERSION "2.8.7"
#define REDIS_VERSION "2.8.10"
+2 -2
View File
@@ -321,8 +321,8 @@ size_t zmalloc_get_rss(void) {
#endif
/* Fragmentation = RSS / allocated-bytes */
float zmalloc_get_fragmentation_ratio(void) {
return (float)zmalloc_get_rss()/zmalloc_used_memory();
float zmalloc_get_fragmentation_ratio(size_t rss) {
return (float)rss/zmalloc_used_memory();
}
#if defined(HAVE_PROC_SMAPS)
+1 -1
View File
@@ -73,7 +73,7 @@ char *zstrdup(const char *s);
size_t zmalloc_used_memory(void);
void zmalloc_enable_thread_safeness(void);
void zmalloc_set_oom_handler(void (*oom_handler)(size_t));
float zmalloc_get_fragmentation_ratio(void);
float zmalloc_get_fragmentation_ratio(size_t rss);
size_t zmalloc_get_rss(void);
size_t zmalloc_get_private_dirty(void);
void zlibc_free(void *ptr);
+60
View File
@@ -6,6 +6,66 @@ start_server {tags {"repl"}} {
s -1 role
} {slave}
test {If min-slaves-to-write is honored, write is accepted} {
r config set min-slaves-to-write 1
r config set min-slaves-max-lag 10
r set foo 12345
wait_for_condition 50 100 {
[r -1 get foo] eq {12345}
} else {
fail "Write did not reached slave"
}
}
test {No write if min-slaves-to-write is < attached slaves} {
r config set min-slaves-to-write 2
r config set min-slaves-max-lag 10
catch {r set foo 12345} err
set err
} {NOREPLICAS*}
test {If min-slaves-to-write is honored, write is accepted (again)} {
r config set min-slaves-to-write 1
r config set min-slaves-max-lag 10
r set foo 12345
wait_for_condition 50 100 {
[r -1 get foo] eq {12345}
} else {
fail "Write did not reached slave"
}
}
test {No write if min-slaves-max-lag is > of the slave lag} {
r -1 deferred 1
r config set min-slaves-to-write 1
r config set min-slaves-max-lag 2
r -1 debug sleep 6
assert {[r set foo 12345] eq {OK}}
after 4000
catch {r set foo 12345} err
assert {[r -1 read] eq {OK}}
r -1 deferred 0
set err
} {NOREPLICAS*}
test {min-slaves-to-write is ignored by slaves} {
r config set min-slaves-to-write 1
r config set min-slaves-max-lag 10
r -1 config set min-slaves-to-write 1
r -1 config set min-slaves-max-lag 10
r set foo aaabbb
wait_for_condition 50 100 {
[r -1 get foo] eq {aaabbb}
} else {
fail "Write did not reached slave"
}
}
# Fix parameters for the next test to work
r config set min-slaves-to-write 0
r -1 config set min-slaves-to-write 0
r flushall
test {MASTER and SLAVE dataset should be identical after complex ops} {
createComplexDataset r 10000
after 500
+53 -15
View File
@@ -13,6 +13,7 @@ proc 03_test_slaves_replication {} {
set master_port [RI $master_id tcp_port]
foreach_redis_id id {
if {$id == $master_id} continue
if {[instance_is_killed redis $id]} continue
wait_for_condition 1000 50 {
[RI $id master_port] == $master_port
} else {
@@ -23,23 +24,60 @@ proc 03_test_slaves_replication {} {
}
}
03_test_slaves_replication
test "Crash the master and force a failover" {
set old_port [RI $master_id tcp_port]
set addr [S 0 SENTINEL GET-MASTER-ADDR-BY-NAME mymaster]
assert {[lindex $addr 1] == $old_port}
kill_instance redis $master_id
foreach_sentinel_id id {
wait_for_condition 1000 50 {
[lindex [S $id SENTINEL GET-MASTER-ADDR-BY-NAME mymaster] 1] != $old_port
} else {
fail "At least one Sentinel did not received failover info"
proc 03_crash_and_failover {} {
uplevel 1 {
test "Crash the master and force a failover" {
set old_port [RI $master_id tcp_port]
set addr [S 0 SENTINEL GET-MASTER-ADDR-BY-NAME mymaster]
assert {[lindex $addr 1] == $old_port}
kill_instance redis $master_id
foreach_sentinel_id id {
wait_for_condition 1000 50 {
[lindex [S $id SENTINEL GET-MASTER-ADDR-BY-NAME mymaster] 1] != $old_port
} else {
fail "At least one Sentinel did not received failover info"
}
}
restart_instance redis $master_id
set addr [S 0 SENTINEL GET-MASTER-ADDR-BY-NAME mymaster]
set master_id [get_instance_id_by_port redis [lindex $addr 1]]
}
}
restart_instance redis $master_id
set addr [S 0 SENTINEL GET-MASTER-ADDR-BY-NAME mymaster]
set master_id [get_instance_id_by_port redis [lindex $addr 1]]
}
03_test_slaves_replication
03_crash_and_failover
03_test_slaves_replication
test "Kill a slave instance" {
foreach_redis_id id {
if {$id == $master_id} continue
set killed_slave_id $id
kill_instance redis $id
break
}
}
03_crash_and_failover
03_test_slaves_replication
test "Wait for failover to end" {
set inprogress 1
while {$inprogress} {
set inprogress 0
foreach_sentinel_id id {
if {[dict exists [S $id SENTINEL MASTER mymaster] failover-state]} {
incr inprogress
}
}
if {$inprogress} {after 100}
}
}
test "Restart killed slave and test replication of slaves again..." {
restart_instance redis $killed_slave_id
}
# Now we check if the slave rejoining the partition is reconfigured even
# if the failover finished.
03_test_slaves_replication
+8
View File
@@ -275,6 +275,8 @@ proc foreach_instance_id {instances idvar code} {
error $result $::errorInfo $::errorCode
} elseif {$errcode == 4} {
continue
} elseif {$errcode == 3} {
break
} elseif {$errcode != 0} {
return -code $errcode $result
}
@@ -356,6 +358,12 @@ proc kill_instance {type id} {
set ::pids [lsearch -all -inline -not -exact $::pids $pid]
}
# Return true of the instance of the specified type/id is killed.
proc instance_is_killed {type id} {
set pid [get_instance_attrib $type $id pid]
return $pid == -1
}
# Restart an instance previously killed by kill_instance
proc restart_instance {type id} {
set dirname "${type}_${id}"
+5
View File
@@ -40,6 +40,11 @@ proc kill_server config {
test "Check for memory leaks (pid $pid)" {
set output {0 leaks}
catch {exec leaks $pid} output
if {[string match {*process does not exist*} $output] ||
[string match {*cannot examine*} $output]} {
# In a few tests we kill the server process.
set output "0 leaks"
}
set output
} {*0 leaks*}
}
+1
View File
@@ -47,6 +47,7 @@ set ::all_tests {
unit/dump
unit/bitops
unit/memefficiency
unit/hyperloglog
}
# Index to the next test to run in the ::all_tests list.
set ::next_test 0
+8
View File
@@ -261,6 +261,14 @@ start_server {tags {"basic"}} {
assert_equal 20 [r get x]
}
test "DEL against expired key" {
r debug set-active-expire 0
r setex keyExpire 1 valExpire
after 1100
assert_equal 0 [r del keyExpire]
r debug set-active-expire 1
}
test {EXISTS} {
set res {}
r set newkey test
+170
View File
@@ -0,0 +1,170 @@
start_server {tags {"hll"}} {
test {HyperLogLog self test passes} {
catch {r pfselftest} e
set e
} {OK}
test {PFADD without arguments creates an HLL value} {
r pfadd hll
r exists hll
} {1}
test {Approximated cardinality after creation is zero} {
r pfcount hll
} {0}
test {PFADD returns 1 when at least 1 reg was modified} {
r pfadd hll a b c
} {1}
test {PFADD returns 0 when no reg was modified} {
r pfadd hll a b c
} {0}
test {PFADD works with empty string (regression)} {
r pfadd hll ""
}
# Note that the self test stresses much better the
# cardinality estimation error. We are testing just the
# command implementation itself here.
test {PFCOUNT returns approximated cardinality of set} {
r del hll
set res {}
r pfadd hll 1 2 3 4 5
lappend res [r pfcount hll]
# Call it again to test cached value invalidation.
r pfadd hll 6 7 8 8 9 10
lappend res [r pfcount hll]
set res
} {5 10}
test {HyperLogLogs are promote from sparse to dense} {
r del hll
r config set hll-sparse-max-bytes 3000
set n 0
while {$n < 100000} {
set elements {}
for {set j 0} {$j < 100} {incr j} {lappend elements [expr rand()]}
incr n 100
r pfadd hll {*}$elements
set card [r pfcount hll]
set err [expr {abs($card-$n)}]
assert {$err < (double($card)/100)*5}
if {$n < 1000} {
assert {[r pfdebug encoding hll] eq {sparse}}
} elseif {$n > 10000} {
assert {[r pfdebug encoding hll] eq {dense}}
}
}
}
test {HyperLogLog sparse encoding stress test} {
for {set x 0} {$x < 1000} {incr x} {
r del hll1 hll2
set numele [randomInt 100]
set elements {}
for {set j 0} {$j < $numele} {incr j} {
lappend elements [expr rand()]
}
# Force dense representation of hll2
r pfadd hll2
r pfdebug todense hll2
r pfadd hll1 {*}$elements
r pfadd hll2 {*}$elements
assert {[r pfdebug encoding hll1] eq {sparse}}
assert {[r pfdebug encoding hll2] eq {dense}}
# Cardinality estimated should match exactly.
assert {[r pfcount hll1] eq [r pfcount hll2]}
}
}
test {Corrupted sparse HyperLogLogs are detected: Additionl at tail} {
r del hll
r pfadd hll a b c
r append hll "hello"
set e {}
catch {r pfcount hll} e
set e
} {*INVALIDOBJ*}
test {Corrupted sparse HyperLogLogs are detected: Broken magic} {
r del hll
r pfadd hll a b c
r setrange hll 0 "0123"
set e {}
catch {r pfcount hll} e
set e
} {*WRONGTYPE*}
test {Corrupted sparse HyperLogLogs are detected: Invalid encoding} {
r del hll
r pfadd hll a b c
r setrange hll 4 "x"
set e {}
catch {r pfcount hll} e
set e
} {*WRONGTYPE*}
test {Corrupted dense HyperLogLogs are detected: Wrong length} {
r del hll
r pfadd hll a b c
r setrange hll 4 "\x00"
set e {}
catch {r pfcount hll} e
set e
} {*WRONGTYPE*}
test {PFADD, PFCOUNT, PFMERGE type checking works} {
r set foo bar
catch {r pfadd foo 1} e
assert_match {*WRONGTYPE*} $e
catch {r pfcount foo} e
assert_match {*WRONGTYPE*} $e
catch {r pfmerge bar foo} e
assert_match {*WRONGTYPE*} $e
catch {r pfmerge foo bar} e
assert_match {*WRONGTYPE*} $e
}
test {PFMERGE results on the cardinality of union of sets} {
r del hll hll1 hll2 hll3
r pfadd hll1 a b c
r pfadd hll2 b c d
r pfadd hll3 c d e
r pfmerge hll hll1 hll2 hll3
r pfcount hll
} {5}
test {PFCOUNT multiple-keys merge returns cardinality of union} {
r del hll1 hll2 hll3
for {set x 1} {$x < 10000} {incr x} {
# Force dense representation of hll2
r pfadd hll1 "foo-$x"
r pfadd hll2 "bar-$x"
r pfadd hll3 "zap-$x"
set card [r pfcount hll1 hll2 hll3]
set realcard [expr {$x*3}]
set err [expr {abs($card-$realcard)}]
assert {$err < (double($card)/100)*5}
}
}
test {PFDEBUG GETREG returns the HyperLogLog raw registers} {
r del hll
r pfadd hll 1 2 3
llength [r pfdebug getreg hll]
} {16384}
test {PFADD / PFCOUNT cache invalidation works} {
r del hll
r pfadd hll a b c
r pfcount hll
assert {[r getrange hll 15 15] eq "\x00"}
r pfadd hll a b c
assert {[r getrange hll 15 15] eq "\x00"}
r pfadd hll 1 2 3
assert {[r getrange hll 15 15] eq "\x80"}
}
}
+36 -1
View File
@@ -311,8 +311,42 @@ start_server {tags {"scripting"}} {
r config set slave-read-only yes
r slaveof 127.0.0.1 0
r debug loadaof
r get foo
set res [r get foo]
r slaveof no one
set res
} {102}
test {We can call scripts rewriting client->argv from Lua} {
r del myset
r sadd myset a b c
r mset a 1 b 2 c 3 d 4
assert {[r spop myset] ne {}}
assert {[r spop myset] ne {}}
assert {[r spop myset] ne {}}
assert {[r mget a b c d] eq {1 2 3 4}}
assert {[r spop myset] eq {}}
}
test {Call Redis command with many args from Lua (issue #1764)} {
r eval {
local i
local x={}
redis.call('del','mylist')
for i=1,100 do
table.insert(x,i)
end
redis.call('rpush','mylist',unpack(x))
return redis.call('lrange','mylist',0,-1)
} 0
} {1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100}
test {Number conversion precision test (issue #1118)} {
r eval {
local value = 9007199254740991
redis.call("set","foo",value)
return redis.call("get","foo")
} 0
} {9007199254740991}
}
# Start a new server since the last test in this stanza will kill the
@@ -326,6 +360,7 @@ start_server {tags {"scripting"}} {
catch {r ping} e
assert_match {BUSY*} $e
r script kill
after 200 ; # Give some time to Lua to call the hook again...
assert_equal [r ping] "PONG"
}
+161
View File
@@ -296,6 +296,62 @@ start_server {tags {"zset"}} {
assert_error "*not*float*" {r zrangebyscore fooz 1 NaN}
}
proc create_default_lex_zset {} {
create_zset zset {0 alpha 0 bar 0 cool 0 down
0 elephant 0 foo 0 great 0 hill
0 omega}
}
test "ZRANGEBYLEX/ZREVRANGEBYLEX/ZCOUNT basics" {
create_default_lex_zset
# inclusive range
assert_equal {alpha bar cool} [r zrangebylex zset - \[cool]
assert_equal {bar cool down} [r zrangebylex zset \[bar \[down]
assert_equal {great hill omega} [r zrangebylex zset \[g +]
assert_equal {cool bar alpha} [r zrevrangebylex zset \[cool -]
assert_equal {down cool bar} [r zrevrangebylex zset \[down \[bar]
assert_equal {omega hill great foo elephant down} [r zrevrangebylex zset + \[d]
assert_equal 3 [r zlexcount zset \[ele \[h]
# exclusive range
assert_equal {alpha bar} [r zrangebylex zset - (cool]
assert_equal {cool} [r zrangebylex zset (bar (down]
assert_equal {hill omega} [r zrangebylex zset (great +]
assert_equal {bar alpha} [r zrevrangebylex zset (cool -]
assert_equal {cool} [r zrevrangebylex zset (down (bar]
assert_equal {omega hill} [r zrevrangebylex zset + (great]
assert_equal 2 [r zlexcount zset (ele (great]
# inclusive and exclusive
assert_equal {} [r zrangebylex zset (az (b]
assert_equal {} [r zrangebylex zset (z +]
assert_equal {} [r zrangebylex zset - \[aaaa]
assert_equal {} [r zrevrangebylex zset \[elez \[elex]
assert_equal {} [r zrevrangebylex zset (hill (omega]
}
test "ZRANGEBYSLEX with LIMIT" {
create_default_lex_zset
assert_equal {alpha bar} [r zrangebylex zset - \[cool LIMIT 0 2]
assert_equal {bar cool} [r zrangebylex zset - \[cool LIMIT 1 2]
assert_equal {} [r zrangebylex zset \[bar \[down LIMIT 0 0]
assert_equal {} [r zrangebylex zset \[bar \[down LIMIT 2 0]
assert_equal {bar} [r zrangebylex zset \[bar \[down LIMIT 0 1]
assert_equal {cool} [r zrangebylex zset \[bar \[down LIMIT 1 1]
assert_equal {bar cool down} [r zrangebylex zset \[bar \[down LIMIT 0 100]
assert_equal {omega hill great foo elephant} [r zrevrangebylex zset + \[d LIMIT 0 5]
assert_equal {omega hill great foo} [r zrevrangebylex zset + \[d LIMIT 0 4]
}
test "ZRANGEBYLEX with invalid lex range specifiers" {
assert_error "*not*string*" {r zrangebylex fooz foo bar}
assert_error "*not*string*" {r zrangebylex fooz \[foo bar}
assert_error "*not*string*" {r zrangebylex fooz foo \[bar}
assert_error "*not*string*" {r zrangebylex fooz +x \[bar}
assert_error "*not*string*" {r zrangebylex fooz -x \[bar}
}
test "ZREMRANGEBYSCORE basics" {
proc remrangebyscore {min max} {
create_zset zset {1 a 2 b 3 c 4 d 5 e}
@@ -708,6 +764,111 @@ start_server {tags {"zset"}} {
assert_equal {} $err
}
test "ZRANGEBYLEX fuzzy test, 100 ranges in $elements element sorted set - $encoding" {
set lexset {}
r del zset
for {set j 0} {$j < $elements} {incr j} {
set e [randstring 0 30 alpha]
lappend lexset $e
r zadd zset 0 $e
}
set lexset [lsort -unique $lexset]
for {set j 0} {$j < 100} {incr j} {
set min [randstring 0 30 alpha]
set max [randstring 0 30 alpha]
set mininc [randomInt 2]
set maxinc [randomInt 2]
if {$mininc} {set cmin "\[$min"} else {set cmin "($min"}
if {$maxinc} {set cmax "\[$max"} else {set cmax "($max"}
set rev [randomInt 2]
if {$rev} {
set cmd zrevrangebylex
} else {
set cmd zrangebylex
}
# Make sure data is the same in both sides
assert {[r zrange zset 0 -1] eq $lexset}
# Get the Redis output
set output [r $cmd zset $cmin $cmax]
if {$rev} {
set outlen [r zlexcount zset $cmax $cmin]
} else {
set outlen [r zlexcount zset $cmin $cmax]
}
# Compute the same output via Tcl
set o {}
set copy $lexset
if {(!$rev && [string compare $min $max] > 0) ||
($rev && [string compare $max $min] > 0)} {
# Empty output when ranges are inverted.
} else {
if {$rev} {
# Invert the Tcl array using Redis itself.
set copy [r zrevrange zset 0 -1]
# Invert min / max as well
lassign [list $min $max $mininc $maxinc] \
max min maxinc mininc
}
foreach e $copy {
set mincmp [string compare $e $min]
set maxcmp [string compare $e $max]
if {
($mininc && $mincmp >= 0 || !$mininc && $mincmp > 0)
&&
($maxinc && $maxcmp <= 0 || !$maxinc && $maxcmp < 0)
} {
lappend o $e
}
}
}
assert {$o eq $output}
assert {$outlen eq [llength $output]}
}
}
test "ZREMRANGEBYLEX fuzzy test, 100 ranges in $elements element sorted set - $encoding" {
set lexset {}
r del zset zsetcopy
for {set j 0} {$j < $elements} {incr j} {
set e [randstring 0 30 alpha]
lappend lexset $e
r zadd zset 0 $e
}
set lexset [lsort -unique $lexset]
for {set j 0} {$j < 100} {incr j} {
# Copy...
r zunionstore zsetcopy 1 zset
set lexsetcopy $lexset
set min [randstring 0 30 alpha]
set max [randstring 0 30 alpha]
set mininc [randomInt 2]
set maxinc [randomInt 2]
if {$mininc} {set cmin "\[$min"} else {set cmin "($min"}
if {$maxinc} {set cmax "\[$max"} else {set cmax "($max"}
# Make sure data is the same in both sides
assert {[r zrange zset 0 -1] eq $lexset}
# Get the range we are going to remove
set torem [r zrangebylex zset $cmin $cmax]
set toremlen [r zlexcount zset $cmin $cmax]
r zremrangebylex zsetcopy $cmin $cmax
set output [r zrange zsetcopy 0 -1]
# Remove the range with Tcl from the original list
if {$toremlen} {
set first [lsearch -exact $lexsetcopy [lindex $torem 0]]
set last [expr {$first+$toremlen-1}]
set lexsetcopy [lreplace $lexsetcopy $first $last]
}
assert {$lexsetcopy eq $output}
}
}
test "ZSETs skiplist implementation backlink consistency test - $encoding" {
set diff 0
for {set j 0} {$j < $elements} {incr j} {
+2 -1
View File
@@ -11,7 +11,8 @@ GROUPS = [
"transactions",
"connection",
"server",
"scripting"
"scripting",
"hyperloglog"
].freeze
GROUPS_BY_NAME = Hash[*
+1
View File
@@ -0,0 +1 @@
*.txt
+27
View File
@@ -0,0 +1,27 @@
# hll-err.rb - Copyright (C) 2014 Salvatore Sanfilippo
# BSD license, See the COPYING file for more information.
#
# Check error of HyperLogLog Redis implementation for different set sizes.
require 'rubygems'
require 'redis'
require 'digest/sha1'
r = Redis.new
r.del('hll')
i = 0
while true do
100.times {
elements = []
1000.times {
ele = Digest::SHA1.hexdigest(i.to_s)
elements << ele
i += 1
}
r.pfadd('hll',*elements)
}
approx = r.pfcount('hll')
abs_err = (approx-i).abs
rel_err = 100.to_f*abs_err/i
puts "#{i} vs #{approx}: #{rel_err}%"
end
+88
View File
@@ -0,0 +1,88 @@
# hll-err.rb - Copyright (C) 2014 Salvatore Sanfilippo
# BSD license, See the COPYING file for more information.
#
# This program is suited to output average and maximum errors of
# the Redis HyperLogLog implementation in a format suitable to print
# graphs using gnuplot.
require 'rubygems'
require 'redis'
require 'digest/sha1'
# Generate an array of [cardinality,relative_error] pairs
# in the 0 - max range, with the specified step.
#
# 'r' is the Redis object used to perform the queries.
# 'seed' must be different every time you want a test performed
# with a different set. The function guarantees that if 'seed' is the
# same, exactly the same dataset is used, and when it is different,
# a totally unrelated different data set is used (without any common
# element in practice).
def run_experiment(r,seed,max,step)
r.del('hll')
i = 0
samples = []
step = 1000 if step > 1000
while i < max do
elements = []
step.times {
ele = Digest::SHA1.hexdigest(i.to_s+seed.to_s)
elements << ele
i += 1
}
r.pfadd('hll',*elements)
approx = r.pfcount('hll')
err = approx-i
rel_err = 100.to_f*err/i
samples << [i,rel_err]
end
samples
end
def filter_samples(numsets,max,step,filter)
r = Redis.new
dataset = {}
(0...numsets).each{|i|
dataset[i] = run_experiment(r,i,max,step)
STDERR.puts "Set #{i}"
}
dataset[0].each_with_index{|ele,index|
if filter == :max
card=ele[0]
err=ele[1].abs
(1...numsets).each{|i|
err = dataset[i][index][1] if err < dataset[i][index][1]
}
puts "#{card} #{err}"
elsif filter == :avg
card=ele[0]
err = 0
(0...numsets).each{|i|
err += dataset[i][index][1]
}
err /= numsets
puts "#{card} #{err}"
elsif filter == :absavg
card=ele[0]
err = 0
(0...numsets).each{|i|
err += dataset[i][index][1].abs
}
err /= numsets
puts "#{card} #{err}"
elsif filter == :all
(0...numsets).each{|i|
card,err = dataset[i][index]
puts "#{card} #{err}"
}
else
raise "Unknown filter #{filter}"
end
}
end
if ARGV.length != 4
puts "Usage: hll-gnuplot-graph <samples> <max> <step> (max|avg|absavg|all)"
exit 1
end
filter_samples(ARGV[0].to_i,ARGV[1].to_i,ARGV[2].to_i,ARGV[3].to_sym)
+116 -61
View File
@@ -30,46 +30,48 @@
# this scripts should be run as root
die () {
echo "ERROR: $1. Aborting!"
echo "ERROR: $1. Aborting!"
exit 1
}
#Absolute path to this script
SCRIPT=$(readlink -f $0)
#Absolute path this script is in
SCRIPTPATH=$(dirname $SCRIPT)
#Initial defaults
_REDIS_PORT=6379
echo "Welcome to the redis service installer"
echo "This script will help you easily set up a running redis server
echo "This script will help you easily set up a running redis server"
echo
"
#check for root user TODO: replace this with a call to "id"
if [ `whoami` != "root" ] ; then
#check for root user
if [ "$(id -u)" -ne 0 ] ; then
echo "You must run this script as root. Sorry!"
exit 1
fi
#Read the redis port
read -p "Please select the redis port for this instance: [$_REDIS_PORT] " REDIS_PORT
if [ ! `echo $REDIS_PORT | egrep "^[0-9]+\$"` ] ; then
read -p "Please select the redis port for this instance: [$_REDIS_PORT] " REDIS_PORT
if ! echo $REDIS_PORT | egrep -q '^[0-9]+$' ; then
echo "Selecting default: $_REDIS_PORT"
REDIS_PORT=$_REDIS_PORT
REDIS_PORT=$_REDIS_PORT
fi
#read the redis config file
_REDIS_CONFIG_FILE="/etc/redis/$REDIS_PORT.conf"
read -p "Please select the redis config file name [$_REDIS_CONFIG_FILE] " REDIS_CONFIG_FILE
if [ !"$REDIS_CONFIG_FILE" ] ; then
if [ -z "$REDIS_CONFIG_FILE" ] ; then
REDIS_CONFIG_FILE=$_REDIS_CONFIG_FILE
echo "Selected default - $REDIS_CONFIG_FILE"
fi
#try and create it
mkdir -p `dirname "$REDIS_CONFIG_FILE"` || die "Could not create redis config directory"
#read the redis log file path
_REDIS_LOG_FILE="/var/log/redis_$REDIS_PORT.log"
read -p "Please select the redis log file name [$_REDIS_LOG_FILE] " REDIS_LOG_FILE
if [ !"$REDIS_LOG_FILE" ] ; then
if [ -z "$REDIS_LOG_FILE" ] ; then
REDIS_LOG_FILE=$_REDIS_LOG_FILE
echo "Selected default - $REDIS_LOG_FILE"
fi
@@ -78,55 +80,71 @@ fi
#get the redis data directory
_REDIS_DATA_DIR="/var/lib/redis/$REDIS_PORT"
read -p "Please select the data directory for this instance [$_REDIS_DATA_DIR] " REDIS_DATA_DIR
if [ !"$REDIS_DATA_DIR" ] ; then
if [ -z "$REDIS_DATA_DIR" ] ; then
REDIS_DATA_DIR=$_REDIS_DATA_DIR
echo "Selected default - $REDIS_DATA_DIR"
fi
mkdir -p $REDIS_DATA_DIR || die "Could not create redis data directory"
#get the redis executable path
_REDIS_EXECUTABLE=`which redis-server`
_REDIS_EXECUTABLE=`command -v redis-server`
read -p "Please select the redis executable path [$_REDIS_EXECUTABLE] " REDIS_EXECUTABLE
if [ ! -f "$REDIS_EXECUTABLE" ] ; then
if [ ! -x "$REDIS_EXECUTABLE" ] ; then
REDIS_EXECUTABLE=$_REDIS_EXECUTABLE
if [ ! -f "$REDIS_EXECUTABLE" ] ; then
if [ ! -x "$REDIS_EXECUTABLE" ] ; then
echo "Mmmmm... it seems like you don't have a redis executable. Did you run make install yet?"
exit 1
fi
fi
#render the tmplates
TMP_FILE="/tmp/$REDIS_PORT.conf"
DEFAULT_CONFIG="../redis.conf"
INIT_TPL_FILE="./redis_init_script.tpl"
INIT_SCRIPT_DEST="/etc/init.d/redis_$REDIS_PORT"
PIDFILE="/var/run/redis_$REDIS_PORT.pid"
#check the default for redis cli
CLI_EXEC=`which redis-cli`
if [ ! "$CLI_EXEC" ] ; then
CLI_EXEC=`command -v redis-cli`
if [ -z "$CLI_EXEC" ] ; then
CLI_EXEC=`dirname $REDIS_EXECUTABLE`"/redis-cli"
fi
echo "Selected config:"
echo "Port : $REDIS_PORT"
echo "Config file : $REDIS_CONFIG_FILE"
echo "Log file : $REDIS_LOG_FILE"
echo "Data dir : $REDIS_DATA_DIR"
echo "Executable : $REDIS_EXECUTABLE"
echo "Cli Executable : $CLI_EXEC"
read -p "Is this ok? Then press ENTER to go on or Ctrl-C to abort." _UNUSED_
mkdir -p `dirname "$REDIS_CONFIG_FILE"` || die "Could not create redis config directory"
mkdir -p `dirname "$REDIS_LOG_FILE"` || die "Could not create redis log dir"
mkdir -p "$REDIS_DATA_DIR" || die "Could not create redis data directory"
#render the templates
TMP_FILE="/tmp/${REDIS_PORT}.conf"
DEFAULT_CONFIG="${SCRIPTPATH}/../redis.conf"
INIT_TPL_FILE="${SCRIPTPATH}/redis_init_script.tpl"
INIT_SCRIPT_DEST="/etc/init.d/redis_${REDIS_PORT}"
PIDFILE="/var/run/redis_${REDIS_PORT}.pid"
if [ ! -f "$DEFAULT_CONFIG" ]; then
echo "Mmmmm... the default config is missing. Did you switch to the utils directory?"
exit 1
fi
#Generate config file from the default config file as template
#changing only the stuff we're controlling from this script
echo "## Generated by install_server.sh ##" > $TMP_FILE
SED_EXPR="s#^port [0-9]{4}\$#port ${REDIS_PORT}#;\
s#^logfile .+\$#logfile ${REDIS_LOG_FILE}#;\
s#^dir .+\$#dir ${REDIS_DATA_DIR}#;\
s#^pidfile .+\$#pidfile ${PIDFILE}#;\
s#^daemonize no\$#daemonize yes#;"
echo $SED_EXPR
read -r SED_EXPR <<-EOF
s#^port [0-9]{4}\$#port ${REDIS_PORT}#; \
s#^logfile .+\$#logfile ${REDIS_LOG_FILE}#; \
s#^dir .+\$#dir ${REDIS_DATA_DIR}#; \
s#^pidfile .+\$#pidfile ${PIDFILE}#; \
s#^daemonize no\$#daemonize yes#;
EOF
sed -r "$SED_EXPR" $DEFAULT_CONFIG >> $TMP_FILE
#cat $TPL_FILE | while read line; do eval "echo \"$line\"" >> $TMP_FILE; done
cp -f $TMP_FILE $REDIS_CONFIG_FILE || exit 1
cp $TMP_FILE $REDIS_CONFIG_FILE || die "Could not write redis config file $REDIS_CONFIG_FILE"
#Generate sample script from template file
rm -f $TMP_FILE
@@ -138,7 +156,7 @@ REDIS_INIT_HEADER=\
#Configurations injected by install_server below....\n\n
EXEC=$REDIS_EXECUTABLE\n
CLIEXEC=$CLI_EXEC\n
PIDFILE=$PIDFILE\n
PIDFILE=\"$PIDFILE\"\n
CONF=\"$REDIS_CONFIG_FILE\"\n\n
REDISPORT=\"$REDIS_PORT\"\n\n
###############\n\n"
@@ -146,45 +164,82 @@ REDISPORT=\"$REDIS_PORT\"\n\n
REDIS_CHKCONFIG_INFO=\
"# REDHAT chkconfig header\n\n
# chkconfig: - 58 74\n
# description: redis_6379 is the redis daemon.\n
# description: redis_${REDIS_PORT} is the redis daemon.\n
### BEGIN INIT INFO\n
# Provides: redis_6379\n
# Required-Start: $network $local_fs $remote_fs\n
# Required-Stop: $network $local_fs $remote_fs\n
# Required-Start: \$network \$local_fs \$remote_fs\n
# Required-Stop: \$network \$local_fs \$remote_fs\n
# Default-Start: 2 3 4 5\n
# Default-Stop: 0 1 6\n
# Should-Start: $syslog $named\n
# Should-Stop: $syslog $named\n
# Short-Description: start and stop redis_6379\n
# Should-Start: \$syslog \$named\n
# Should-Stop: \$syslog \$named\n
# Short-Description: start and stop redis_${REDIS_PORT}\n
# Description: Redis daemon\n
### END INIT INFO\n\n"
if [ !`which chkconfig` ] ; then
#combine the header and the template (which is actually a static footer)
echo $REDIS_INIT_HEADER > $TMP_FILE && cat $INIT_TPL_FILE >> $TMP_FILE || die "Could not write init script to $TMP_FILE"
else
if command -v chkconfig >/dev/null; then
#if we're a box with chkconfig on it we want to include info for chkconfig
echo -e $REDIS_INIT_HEADER $REDIS_CHKCONFIG_INFO > $TMP_FILE && cat $INIT_TPL_FILE >> $TMP_FILE || die "Could not write init script to $TMP_FILE"
echo "$REDIS_INIT_HEADER" "$REDIS_CHKCONFIG_INFO" > $TMP_FILE && cat $INIT_TPL_FILE >> $TMP_FILE || die "Could not write init script to $TMP_FILE"
else
#combine the header and the template (which is actually a static footer)
echo "$REDIS_INIT_HEADER" > $TMP_FILE && cat $INIT_TPL_FILE >> $TMP_FILE || die "Could not write init script to $TMP_FILE"
fi
###
# Generate sample script from template file
# - No need to check which system we are on. The init info are comments and
# do not interfere with update_rc.d systems. Additionally:
# Ubuntu/debian by default does not come with chkconfig, but does issue a
# warning if init info is not available.
cat > ${TMP_FILE} <<EOT
#/bin/sh
#Configurations injected by install_server below....
EXEC=$REDIS_EXECUTABLE
CLIEXEC=$CLI_EXEC
PIDFILE=$PIDFILE
CONF="$REDIS_CONFIG_FILE"
REDISPORT="$REDIS_PORT"
###############
# SysV Init Information
# chkconfig: - 58 74
# description: redis_${REDIS_PORT} is the redis daemon.
### BEGIN INIT INFO
# Provides: redis_${REDIS_PORT}
# Required-Start: \$network \$local_fs \$remote_fs
# Required-Stop: \$network \$local_fs \$remote_fs
# Default-Start: 2 3 4 5
# Default-Stop: 0 1 6
# Should-Start: \$syslog \$named
# Should-Stop: \$syslog \$named
# Short-Description: start and stop redis_${REDIS_PORT}
# Description: Redis daemon
### END INIT INFO
EOT
cat ${INIT_TPL_FILE} >> ${TMP_FILE}
#copy to /etc/init.d
cp -f $TMP_FILE $INIT_SCRIPT_DEST && chmod +x $INIT_SCRIPT_DEST || die "Could not copy redis init script to $INIT_SCRIPT_DEST"
cp $TMP_FILE $INIT_SCRIPT_DEST && \
chmod +x $INIT_SCRIPT_DEST || die "Could not copy redis init script to $INIT_SCRIPT_DEST"
echo "Copied $TMP_FILE => $INIT_SCRIPT_DEST"
#Install the service
echo "Installing service..."
if [ !`which chkconfig` ] ; then
#if we're not a chkconfig box assume we're able to use update-rc.d
update-rc.d redis_$REDIS_PORT defaults && echo "Success!"
else
if command -v chkconfig >/dev/null 2>&1; then
# we're chkconfig, so lets add to chkconfig and put in runlevel 345
chkconfig --add redis_$REDIS_PORT && echo "Successfully added to chkconfig!"
chkconfig --level 345 redis_$REDIS_PORT on && echo "Successfully added to runlevels 345!"
chkconfig --add redis_${REDIS_PORT} && echo "Successfully added to chkconfig!"
chkconfig --level 345 redis_${REDIS_PORT} on && echo "Successfully added to runlevels 345!"
elif command -v update-rc.d >/dev/null 2>&1; then
#if we're not a chkconfig box assume we're able to use update-rc.d
update-rc.d redis_${REDIS_PORT} defaults && echo "Success!"
else
echo "No supported init tool found."
fi
/etc/init.d/redis_$REDIS_PORT start || die "Failed starting service..."
#tada
echo "Installation successful!"
exit 0
+26 -14
View File
@@ -3,29 +3,41 @@ case "$1" in
start)
if [ -f $PIDFILE ]
then
echo "$PIDFILE exists, process is already running or crashed"
echo "$PIDFILE exists, process is already running or crashed"
else
echo "Starting Redis server..."
$EXEC $CONF
echo "Starting Redis server..."
$EXEC $CONF
fi
;;
stop)
if [ ! -f $PIDFILE ]
then
echo "$PIDFILE does not exist, process is not running"
echo "$PIDFILE does not exist, process is not running"
else
PID=$(cat $PIDFILE)
echo "Stopping ..."
$CLIEXEC -p $REDISPORT shutdown
while [ -x /proc/${PID} ]
do
echo "Waiting for Redis to shutdown ..."
sleep 1
done
echo "Redis stopped"
PID=$(cat $PIDFILE)
echo "Stopping ..."
$CLIEXEC -p $REDISPORT shutdown
while [ -x /proc/${PID} ]
do
echo "Waiting for Redis to shutdown ..."
sleep 1
done
echo "Redis stopped"
fi
;;
status)
if [ ! -f $PIDFILE ]
then
echo 'Redis is not running'
else
echo "Redis is running ($(<$PIDFILE))"
fi
;;
restart)
$0 stop
$0 start
;;
*)
echo "Please use start or stop as first argument"
echo "Please use start, stop, restart or status as first argument"
;;
esac