During the AOF rewrite process, the parent process needs to accumulate
the new writes in an in-memory buffer: when the child will terminate the
AOF rewriting process this buffer (that ist the difference between the
dataset when the rewrite was started, and the current dataset) is
flushed to the new AOF file.
We used to implement this buffer using an sds.c string, but sds.c has a
2GB limit. Sometimes the dataset can be big enough, the amount of writes
so high, and the rewrite process slow enough that we overflow the 2GB
limit, causing a crash, documented on github by issue #504.
In order to prevent this from happening, this commit introduces a new
system to accumulate writes, implemented by a linked list of blocks of
10 MB each, so that we also avoid paying the reallocation cost.
Note that theoretically modern operating systems may implement realloc()
simply as a remaping of the old pages, thus with very good performances,
see for instance the mremap() syscall on Linux. However this is not
always true, and jemalloc by default avoids doing this because there are
issues with the current implementation of mremap().
For this reason we are using a linked list of blocks instead of a single
block that gets reallocated again and again.
The changes in this commit lacks testing, that will be performed before
merging into the unstable branch. This fix will not enter 2.4 because it
is too invasive. However 2.4 will log a warning when the AOF rewrite
buffer is near to the 2GB limit.
The user @jokea noticed that the following line of code into
replication.c made little sense:
addReplySds(slave,sdsempty());
Investigating a bit I found that this was introduced by commit 6208b3a7
three years ago in the early stages of Redis. The code apparently is not
useful at all, so I'm removing it.
This change will not be backported into 2.4 so that in the rare case
this should introduce a bug, we'll have a chance to detect it into the
development branch. However following the code path it seems like the
code is not useful at all, so the risk is truly small.
This makes the code more readable, it is still not the case to split the
file itself into three different files, but the logical separation
improves the readability especially since new commits are going to
introduce an additional section.
Weeks ago trying to fix an harmless GCC warning I introduced a bug in
the ziplist-encoded implementations of sorted sets.
The bug completely broke zuiNext() iterator, that is used in the
ZINTERSTORE and ZUNIONSTORE implementation, so those two commands are no
longer reliable starting from Redis version 2.4.12 and latest 2.6.0-RC
releases.
This commit fixes the problem and adds a regression test.
redis-cli.c uses the time() function to seed the PRNG, but time.h was
not included. This was not noticed since sys/time.h is included and was
enough in most systems (but not correct). With Ubuntu 12.04 GCC
generates a warning that made us aware of the issue.
activeExpireCycle() can consume no more than a few milliseconds per
iteration. This commit improves the precision of the check for the time
elapsed in two ways:
1) We check every 16 iterations instead of the main loop instead of 256.
2) We reset iterations at the start of the function and not every time
we switch to the next database, so the check is correctly performed
every 16 iterations.
A previous commit introduced REDIS_HZ define that changes the frequency
of calls to the serverCron() Redis function. This commit improves
different related things:
1) Software watchdog: now the minimal period can be set according to
REDIS_HZ. The minimal period is two times the timer period, that is:
(1000/REDIS_HZ)*2 milliseconds
2) The incremental rehashing is now performed in the expires dictionary
as well.
3) The activeExpireCycle() function was improved in different ways:
- Now it checks if it already used too much time using microseconds
instead of milliseconds for better precision.
- The time limit is now calculated correctly, in the previous version
the division was performed before of the multiplication resulting in
a timelimit of 0 if HZ was big enough.
- Databases with less than 1% of buckets fill in the hash table are
skipped, because getting random keys is too expensive in this
condition.
4) tryResizeHashTables() is now called at every timer call, we need to
match the number of calls we do to the expired keys colleciton cycle.
5) REDIS_HZ was raised to 100.
Redis uses a function called serverCron() that is very similar to the
timer interrupt of an operating system. This function is used to handle
a number of asynchronous things, like active expired keys collection,
clients timeouts, update of statistics, things related to the cluster
and replication, triggering of BGSAVE and AOF rewrite process, and so
forth.
In the past the timer was called 1 time per second. At some point it was
raised to 10 times per second, but it still was fixed and could not be
changed even at compile time, because different functions called from
serverCron() assumed a given fixed frequency.
This commmit makes the frequency configurable, so that it is simpler to
pick a good tradeoff between overhead of this function (that is usually
very small) and the responsiveness of Redis during a few critical
circumstances where a lot of work is done inside the timer.
An example of such a critical condition is mass-expire of a lot of keys
in the same second. Up to a given percentage of CPU time is used to
perform expired keys collection per expire cylce. Now changing the
REDIS_HZ macro it is possible to do less work but more times per second
in order to block the server for less time.
If this patch will work well in our tests it will enter Redis 2.6-final.
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.
Because of the introduction of new integer encoding types for ziplists
in the 2.6 tree, the same integer value may have a different encoding in
different versions of the ziplist implementation. This means that the
encoding can NOT be used as a fast path in comparing integers.
The new implementation start reading / writing before blocking with
aeWait(), likely the descriptor can accept writes or has buffered data
inside and we can go faster, otherwise we get an error and wait.
This change has effects on speed but also on correctness: on socket
errors when we perform non blocking connect(2) write is performed ASAP
and the error is returned ASAP before waiting.
So the practical effect is that now a Redis slave is more available if it
can not connect to the master, previously the slave continued to block on
syncWrite() trying to send SYNC, and serving commands very slowly.
We used to reply -ERR ... message ..., now the reply is
instead -MASTERDOWN ... message ... so that it can be distinguished
easily by the other error conditions.
Every matched key in a KEYS call is checked for expiration. When the key
is set to expire, the call to `getExpire` will assert that the key also
exists in the main dictionary. This in turn causes a rehashing step to
be executed. Rehashing a dictionary when there is an iterator active may
result in the iterator emitting duplicate entries, or not emitting some
entries at all. By using a safe iterator, the rehash step is omitted.
An user reported a crash with Redis scripting (see issue #480 on
github), inspection of the kindly provided strack trace showed that
server.lua_caller was probably set to NULL. The stack trace also slowed
that the call to the hook was originating from a point where we just
used to set/get a few global variables in the Lua state.
What was happening is that we did not set the timeout hook selectively
only when the user script was called. Now we set it more selectively,
specifically only in the context of the lua_pcall() call, and make sure
to remove the hook when the call returns. Otherwise the hook can get
called in random contexts every time we do something with the Lua
state.
A previous commit removed -g -rdynamic -ggdb as LDFLAGS, not allowing
Redis to produce a stack trace wth symbol names on crash.
This commit fixes the issue.
This commit reverts most of c575766202773c858be0870c20cd495b722927c3, in
order to use back main stack for signal handling.
The main reason is that otherwise it is completely pointless that we do
a lot of efforts to print the stack trace on crash, and the content of
the stack and registers as well. Using an alternate stack broken this
feature completely.
1) One integer "immediate" encoding that can encode from 0 to 12 in the
encoding byte itself.
2) One 8 bit signed integer encoding that can encode 8 bit signed small
integers in a single byte.
The idea is to exploit all the not used bits we have around in a
backward compatible way.
This fixes compilation on FreeBSD (and possibly other systems) by
not using ucontext_t at all if HAVE_BACKTRACE is not defined.
Also the ifdefs to get the registers are modified to explicitly test for the
operating system in the first level, and the arch in the second level
of nesting.
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.
