Commit Graph

8770 Commits

Author SHA1 Message Date
3a3978b10b WITHSCORES in ZRANGEBYSCORE thanks to Sam Hendley 2010-02-07 10:32:28 +01:00
0500ef2736 Added "withscores" option to zrangebyscore command. Based on withscores support in zrange function, ugliest part was the argument parsing to handle
using it with the limit option.
2010-02-06 18:41:25 -05:00
59146ef3c1 DEBUG OBJECT provide info about serialized object length even when VM is disabled 2010-02-06 18:53:39 +01:00
6766f45ef2 multi bulk requests in redis-benchmark, default fsync policy changed to everysec, added a prefix character for DEBUG logs 2010-02-06 13:39:07 +01:00
3c290b9b52 APPEND tests 2010-02-04 16:38:33 +01:00
4b00bebd3d APPEND command 2010-02-04 15:55:52 +01:00
ed9e496634 Faster version of the function hashing possibly encoded objects, leading to a general speed gain when working with Sets of integers 2010-02-02 12:19:24 +01:00
3c68de9b01 faster Set loading time from .rdb file resizing the hash table to the right size before loading elements 2010-02-02 12:05:15 +01:00
9651a78787 Log time taken to load the DB at startup, in seconds 2010-02-02 12:01:53 +01:00
478c2c6f5b Fixed VM corruption due to child fclosing the VM file directly or indirectly calling exit(), now replaced with _exit() in all the sensible places. Masked a few signals from IO threads. 2010-01-31 16:03:44 +01:00
d5d55fc319 loading side of the threaded VM 2010-01-28 10:12:04 -05:00
a544018d04 TODO cahnges 2010-01-26 05:02:27 -05:00
b72f6a4b70 Fixed memory human style memory reporting, removed server.usedmemory, now zmalloc_used_memory() is used always. vm-playpen 2010-01-23 11:55:04 -05:00
b0d8747dae VM tuning thanks to redis-stat vmstat. Now it performs much better under high load 2010-01-22 12:58:11 -05:00
eb6845621c Changelog updated 2010-01-21 10:36:32 -05:00
f6c0bba8bc REDIS_MAX_COMPLETED_JOBS_PROCESSED is now in percentage, not number of jobs. Moved a debugging message a few lines forward as it was called where a few logged parameters where invalid, leading to a crash 2010-01-21 07:27:32 -05:00
76b7233adb fixed a deadlock caused by too much finished processes in queue so that I/O clients writing to the wirte side of the pipe used to awake the main thread where blocking. Then a BGSAVE started waiting for the last active thread to finish, condition impossible because all the I/O threads where blocking on threads. Takes this as a note to myself... 2010-01-20 16:57:00 -05:00
621d5c1985 ae.c event loop does no longer support exception notifications, as they are fully pointless. Also a theoretical bug that never happens in practice fixed. 2010-01-20 13:38:59 -05:00
8e688711a7 commercial tools stuff removed from the Redis makefile. cotools are now migrated into a different repos 2010-01-19 13:32:36 -05:00
970e10bb34 removed a bug in the function to cancel an I/O job 2010-01-19 13:02:02 -05:00
33843a90aa static symbols update 2010-01-17 11:24:37 -05:00
76a41fb163 removed support for REDIS_HELGRIND_FRIENDLY since Helgrind 3.5.0 is friendly enough even with many threads created and destroyed 2010-01-16 07:57:32 -05:00
a9158272c6 now redis-cli understands -h 2010-01-15 16:42:29 -05:00
6fa987e390 Create swap file only if not exists 2010-01-15 14:02:18 -05:00
459f52a8aa I hate warnings 2010-01-15 13:55:52 -05:00
fefed59798 fixed a minor memory leak in configuration file parsing 2010-01-15 13:35:13 -05:00
2316bb3b42 minor fix 2010-01-15 13:17:43 -05:00
054e426dbd support for named VM swap file. Fixed a few important interaction issues between the background saving processes and IO threads 2010-01-15 13:14:23 -05:00
5f8e5d7c9e fix for the just added new test 2010-01-15 10:37:37 -05:00
02fcfc1e39 useless debugging messages removed 2010-01-15 09:39:44 -05:00
7276646263 new test added 2010-01-15 09:37:26 -05:00
4ad37480f3 thread safe zmalloc used memory counter 2010-01-15 08:52:20 -05:00
b04a5df979 A define to make Redis more helgrind friendly 2010-01-15 05:16:12 -05:00
1064ef8764 removed a few races from threaded VM 2010-01-15 05:09:18 -05:00
2e111efe5a Fixed a never experienced, theoretical bug that can actually happen in practice. Basically when a thread is working on a I/O Job we need to wait it to finish before to cancel the Job in vmCancelThreadedIOJob(), otherwise the thread may mess with an object that is being manipulated by the main thread as well. 2010-01-14 17:18:27 -05:00
bcaa7a4ffa Set the new threads stack size to a LZF friendly amount 2010-01-14 08:58:41 -05:00
427a2153d3 access to already freed job structure fixed by statements reoredering 2010-01-13 16:26:41 -05:00
85a2775298 removed a useless debugging message 2010-01-13 13:41:38 -05:00
4ee9488d7e Wait zero active threads condition before to fork() for BGSAVE or BGREWRITEAOF 2010-01-13 13:38:30 -05:00
c7df85a484 list API is now thread safe 2010-01-13 13:18:07 -05:00
b74880b4ed minor TODO and debugging info changes 2010-01-13 09:40:50 -05:00
72e9fd40b6 support for blocking VM in config file 2010-01-12 15:23:52 -05:00
f2d9f50f6c more non blocking VM changes 2010-01-12 15:09:11 -05:00
d663729a12 fix for test #11 2010-01-12 10:01:59 -05:00
25fd2cb284 a few more stuff in INFO about VM. Test #11 changed a bit in order to be less lame 2010-01-12 09:57:00 -05:00
c953f24bce Added a define to configure how many completed IO jobs the handler should process at every call. 2010-01-12 07:27:54 -05:00
6c96ba7dfe Fixed a bug in the IO Job canceling funtion 2010-01-11 19:47:18 -05:00
a5819310b0 more steps towards a working non blocking VM 2010-01-11 17:26:58 -05:00
20f5b38867 converted random printfs in debug logs 2010-01-11 13:10:43 -05:00
b3e3d0d78e removed a bug introduced with non blocking VM 2010-01-11 12:54:29 -05:00