antirez [Thu, 15 Mar 2012 19:51:35 +0000 (20:51 +0100)]
Fix for issue #391.
Use a simple protocol between clientsCron() and helper functions to
understand if the client is still valind and clientsCron() should
continue processing or if the client was freed and we should continue
with the next one.
antirez [Wed, 14 Mar 2012 09:13:23 +0000 (10:13 +0100)]
sds.c new function sdsRemoveFreeSpace().
The new function is used in order to resize the string allocation so
that only the minimal allocation possible is used, removing all the free
space at the end of the string normally used to improve efficiency of
concatenation operations.
antirez [Tue, 13 Mar 2012 09:59:29 +0000 (10:59 +0100)]
RDB hashes loading, fixed another bug in the loading of HT-encoded hashes: when the hash entry is too big for ziplist, add the field, then convert. The code used to break before the new entry was inserted, resulting into missing fields in the loaded Hash object.
antirez [Sat, 10 Mar 2012 09:38:53 +0000 (10:38 +0100)]
Removed handling of deprecated hash-max-zipmap-entries nad hash-map-zipmap-value. Pieter is too good with users ;). Better to have them switch to a saner configuration ASAP after the 2.6 upgrade.
antirez [Thu, 8 Mar 2012 09:13:12 +0000 (10:13 +0100)]
run_id added to INFO output.
The Run ID is a field that identifies a single execution of the Redis
server. It can be useful for many purposes as it makes easy to detect if
the instance we are talking about is the same, or if it is a different
one or was rebooted. An application of run_id will be in the partial
synchronization of replication, where a slave may request a partial sync
from a given offset only if it is talking with the same master. Another
application is in failover and monitoring scripts.
antirez [Wed, 7 Mar 2012 17:02:26 +0000 (18:02 +0100)]
By default Redis refuses writes with an error if the latest BGSAVE failed (and at least one save point is configured). However people having good monitoring systems may prefer a server that continues to work, since they are notified that there are problems by their monitoring systems. This commit implements the ability to turn the feature on or off via redis.conf and CONFIG SET.
antirez [Wed, 7 Mar 2012 12:05:46 +0000 (13:05 +0100)]
Refuse writes if can't persist on disk.
Redis now refuses accepting write queries if RDB persistence is
configured, but RDB snapshots can't be generated for some reason.
The status of the latest background save operation is now exposed
in the INFO output as well. This fixes issue #90.
antirez [Tue, 28 Feb 2012 23:46:50 +0000 (00:46 +0100)]
Better system for additional commands replication.
The new code uses a more generic data structure to describe redis operations.
The new design allows for multiple alsoPropagate() calls within the scope of a
single command, that is useful in different contexts. For instance there
when there are multiple clients doing BRPOPLPUSH against the same list,
and a variadic LPUSH is performed against this list, the blocked clients
will both be served, and we should correctly replicate multiple LPUSH
commands after the replication of the current command.
antirez [Tue, 28 Feb 2012 17:03:08 +0000 (18:03 +0100)]
Added a new API to replicate an additional command after the replication of the currently executed command, in order to propagte the LPUSH originating from RPOPLPUSH and indirectly by BRPOPLPUSH.
antirez [Wed, 22 Feb 2012 08:52:10 +0000 (09:52 +0100)]
zzlIsInRange() now is capable of handling empty sorted sets that may end inside the data set when loading very old RDB files produced by early-stage versions of Redis.
Pieter Noordhuis [Sat, 14 Jan 2012 01:49:16 +0000 (17:49 -0800)]
Don't expire keys when loading an RDB after a SYNC
The cron is responsible for expiring keys. When keys are expired at
load time, it is possible that the snapshot of a master node gets
modified. This can in turn lead to inconsistencies in the data set.
A more concrete example of this behavior follows. A user reported a
slave that would show an monotonically increase input buffer length,
shortly after completing a SYNC. Also, `INFO` output showed a single
blocked client, which could only be the master link. Investigation
showed that indeed the `BRPOP` command was fed by the master. This
command can only end up in the stream of write operations when it did
NOT block, and effectively executed `RPOP`. However, when the key
involved in the `BRPOP` is expired BEFORE the command is executed, the
client executing it will block. The client in this case, is the master
link.
antirez [Tue, 14 Feb 2012 15:15:20 +0000 (16:15 +0100)]
add -f flag to cp when installing, otherwise stopping the server is
needed when installing a new Redis version. Thanks to Scott Kevill.
Fixes issue #335.
antirez [Sat, 4 Feb 2012 13:05:54 +0000 (14:05 +0100)]
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.