9419eb0ffd
More incremental active expired keys collection process.
...
If a large amonut of keys are all expiring about at the same time, the
"active" expired keys collection cycle used to block as far as the
percentage of already expired keys was >= 25% of the total population of
keys with an expire set.
This could block the server even for many seconds in order to reclaim
memory ASAP. The new algorithm uses at max a small amount of
milliseconds per cycle, even if this means reclaiming the memory less
promptly it also means a more responsive server.
2012-05-12 09:33:47 +02:00
df10c797e0
Limit memory used by big SLOWLOG entries.
...
Two limits are added:
1) Up to SLOWLOG_ENTRY_MAX_ARGV arguments are logged.
2) Up to SLOWLOG_ENTRY_MAX_STRING bytes per argument are logged.
3) slowlog-max-len is set to 128 by default (was 1024).
The number of remaining arguments / bytes is logged in the entry
so that the user can understand better the nature of the logged command.
2012-04-21 21:50:25 +02:00
3c38b0876e
Merge pull request #417 from josephjang/e95ca1db21b2dbe82f39a3f7f6390cb5f58dafae
...
a memory leak with replication on two or more dbs including >= db10
2012-03-30 01:28:58 -07:00
e95ca1db21
Fixed a memory leak with replication
...
occurs when two or more dbs are replicated and at least one of them is >db10
2012-03-30 02:17:21 +09:00
6938960b66
Fix for slaves chains. Force resync of slaves (simply disconnecting them) when SLAVEOF turns a master into a slave.
2012-03-29 09:31:39 +02:00
b3a86b8209
Fixed a few broken stuff introduced while merging issue #327 related code in 2.4
2012-02-15 15:30:29 +01:00
35a4761f1d
freeMemoryIfNeeded() minor refactoring
2012-02-15 15:22:57 +01:00
c63e1b83c4
This fixes issue #327 , is a very complex fix (unfortunately), details:
...
1) sendReplyToClient() now no longer stops transferring data to a single
client in the case we are out of memory (maxmemory-wise).
2) in processCommand() the idea of we being out of memory is no longer
the naive zmalloc_used_memory() > server.maxmemory. To say if we can
accept or not write queries is up to the return value of
freeMemoryIfNeeded(), that has full control about that.
3) freeMemoryIfNeeded() now does its math without considering output
buffers size. But at the same time it can't let the output buffers to
put us too much outside the max memory limit, so at the same time it
makes sure there is enough effort into delivering the output buffers to
the slaves, calling the write handler directly.
This three changes are the result of many tests, I found (partially
empirically) that is the best way to address the problem, but maybe
we'll find better solutions in the future.
2012-02-15 15:20:20 +01:00
05755f5c40
Initial version of c->reply_bytes implementation backported from unstable to 2.4, in order to apply issue 327 patches.
2012-02-15 15:20:05 +01:00
f373061563
Added a server.arch_bits field instead of computing it at runtime for INFO.
2012-02-02 10:02:40 +01:00
fbfe656236
On crash print information about the current client (if any), command vector, and object associated to first argument assuming it is a key.
2012-01-12 16:16:23 +01:00
e7b85b3315
Protections against protocol desyncs, leading to infinite query buffer growing, due to nul-terms in specific bytes of the request or indefinitely long multi bulk or bulk count strings without newlines. This bug is related to Issue #141 as well.
2012-01-07 12:50:44 +01:00
a5045d552c
Fixed replication when multiple slaves are attaching at the same time. The output buffer was not copied correctly between slaves. This fixes issue #141 .
2012-01-07 12:48:10 +01:00
8a82ee0f3c
better bug report info on crash (backported from unstable)
2011-11-24 15:53:35 +01:00
f4e2abfcd4
minor refactoring to networking.c adding a separated function to get a string representing the current state of all the connected clients.
2011-11-24 15:47:55 +01:00
3852e2a831
last executed command in CLIENT LIST output.
2011-11-24 14:56:43 +01:00
a6e4627438
Close client connection and log the event when the client input buffer reaches 1GB.
2011-11-21 18:37:17 +01:00
b1b98c64cc
Max log message length set to 4k to avoid truncated INFO output in logs on Redis crash. This is just a workaround and it is handled better in the unstable branch.
2011-11-21 18:32:08 +01:00
beb4bacce6
I/O buffer length enlarged
2011-11-11 17:19:57 +01:00
3fac86ff1d
set default client timeout to zero inside redis.h as well
2011-11-11 17:16:03 +01:00
1fe4cd5de9
fixed source indenting
2011-10-18 19:03:18 -02:00
7c6da7327f
Added repl_ping_slave_period and repl_timeout options support
2011-10-18 18:56:10 -02:00
09cefcfbb3
Added a config directive for a Unix socket mask
...
Added a configuration directive to allow a user to specify the
permissions to be granted to the Unix socket file. I followed
the format Pieter and Salvatore discusses in issue #85 (
https://github.com/antirez/redis/issues/85 ).
2011-10-10 22:08:11 +02:00
79b9ec53d1
fixed flushAppendOnlyFile() function prototype to reflect the new implementation, fixing compilation.
2011-09-19 17:43:56 +02:00
4c2d5f0980
postpone the AOF fsync if policy is everysec and there is a background fsync already going.
2011-09-19 17:43:49 +02:00
3a4afb0473
Manual merge after cherry-pick
2011-07-12 10:03:22 +02:00
63aed54a27
Take a pointer to the relevant entry of the command table in the client structure. This is generally a more sounding design, simplifies a few functions prototype, and as a side effect fixes a bug related to the conversion of EXPIRE -1 to DEL: before of this fix Redis tried to convert it into an EXPIREAT in the AOF code, regardless of our rewrite of the command.
2011-07-12 10:00:12 +02:00
19453556d4
dataspace change hook functions backported to 2.4. They are just wrappers at this point but this makes simpler to cherry-pick between branches.
2011-07-12 09:56:41 +02:00
30b345ddbb
more slow log merge, I forgot to resolve a conflict.
2011-07-02 00:27:10 +02:00
99509ab257
slow log merged into 2.4
2011-07-02 00:24:11 +02:00
5d5fe340e0
new INFO filed master_link_down_since_seconds
2011-06-17 16:13:22 +02:00
307868e149
DB API refactoring. The changes were designed together with Pieter Noordhuis.
2011-06-14 16:59:45 +02:00
61751a4dc4
AOF auto rewrite backported to 2.4
2011-06-10 18:31:27 +02:00
27c8f84c66
INFO now contains the time (in usecs) needed to fork() in order to persist. The info is available in the stats section of INFO.
2011-06-08 23:20:11 +02:00
4f8cf6a23c
use the new rewriteClientCommandVector() function for SPOP -> SREM replication translation as well.
2011-05-31 12:15:55 +02:00
afc151d3ba
Make replication faster (biggest gain for small number of slaves)
2011-05-29 17:55:13 -07:00
ecfc0583da
Configurable synchronous I/O timeout
2011-05-22 12:57:56 +02:00
ff8d85885e
Minor changes in non-blocking repl. connect
2011-05-22 12:48:10 +02:00
78c05caf5e
Non-blocking connect with master
2011-05-19 18:53:31 +02:00
deff2338fa
Merge remote branch 'pietern/2.4-misc' into 2.4
2011-05-05 16:50:28 +02:00
20dad91bdf
Tests for string2ll; move isObject* to object.c
2011-04-27 13:26:33 +02:00
91867a5ea2
CLIENT LIST implemented
2011-04-21 17:09:46 +02:00
1c30cad1ab
peak mem in INFO backported from unstable branch
2011-04-21 16:55:27 +02:00
760dec3a6c
VM/direct-saving fixes
2011-04-20 17:48:46 +02:00
1f84b0648b
yet more manual backporting from unstable
2011-04-20 15:57:35 +02:00
c92283770d
added new RDB codes for ziplist encoded lists and intset encodeed sets
2011-04-20 15:41:13 +02:00
3dffb8e21e
save zipmap encoded hashes as blobs. Work in progress.
2011-04-20 15:41:13 +02:00
dfeb899b2d
Merge branch '2.2' into 2.2-zset
...
Conflicts:
src/object.c
2011-04-06 15:46:33 +02:00
93db956e06
make sure that OBJECT ENCODING returns skiplist for sorted sets, and not raw, so that once we will merge specially encoded sorted sets everything will make sense.
2011-04-06 15:38:08 +02:00
cf6c3f4b04
OBJECT command implemented
2011-04-06 12:22:09 +02:00
