i hava got a problem in function "sdsMakeRoomFor"

[monRedis]

when i use hiredis,i get a problem in function sdsMakeRoomFor.this problem does't accre frequently,only sometimes. follow is this coredump partly,why this is cause? can you tell me how to fix it?
#0 0x007c2854 in SLL_Next(void*) () from /usr/local/lib/libtcmalloc.so.0
#1 0x007c28ad in SLL_Pop(void**) () from /usr/local/lib/libtcmalloc.so.0
#2 0x007c70a5 in TCMalloc_ThreadCache_FreeList::Pop() ()

from /usr/local/lib/libtcmalloc.so.0
#3 0x007c8344 in TCMalloc_ThreadCache::Allocate(unsigned int) ()

from /usr/local/lib/libtcmalloc.so.0
#4 0x007c682b in do_malloc(unsigned int) ()

from /usr/local/lib/libtcmalloc.so.0
#5 0x007e60ed in realloc () from /usr/local/lib/libtcmalloc.so.0
#6 0x001371f5 in sdsMakeRoomFor (s=0x0, t=0x8744564, len=6) at sds.c:97
#7 sdscatlen (s=0x0, t=0x8744564, len=6) at sds.c:131
#8 0x00134f72 in redisvFormatCommand (target=0xb7504d9c,

format=0x805a07e "HSET %s %s %s",
ap=0xb7504e18 "dEt\b\354\207\360\t\334\033u\b\377\377\377\377HNP\267\217\f\0                                                                                       05\b`\263\344\b14207") at hiredis.c:738

#9 0x001353bf in redisvAppendCommand (c=0x8755dc0,

format=0x805a07e "HSET %s %s %s",
ap=0xb7504e18 "dEt\b\354\207\360\t\334\033u\b\377\377\377\377HNP\267\217\f\0                                                                                       05\b`\263\344\b14207") at hiredis.c:1199

#10 0x00136fe0 in redisvCommand (c=0x8755dc0,

format=0x805a07e "HSET %s %s %s",
ap=0xb7504e18 "dEt\b\354\207\360\t\334\033u\b\377\377\377\377HNP\267\217\f\0                                                                                       05\b`\263\344\b14207") at hiredis.c:1266

#11 0x0013704e in redisCommand (c=0x8755dc0, format=0x805a07e "HSET %s %s %s")

at hiredis.c:1275

#12 0x08054704 in CRedisImplCommond::ImplCommondHashSet (this=0x874daf4,

key=0x8744564 "DJY001", value=0x8751bdc "namename|123456|online|10")

[goelvivek]

hi,
Any update here ?
I am getting same issue

Program terminated with signal 11, Segmentation fault.
#0 0x00007f71b5d43693 in malloc_consolidate.part.3 () from /lib64/libc.so.6
#0 0x00007f71b5d43693 in malloc_consolidate.part.3 () from /lib64/libc.so.6
#1 0x00007f71b5d44b28 in _int_malloc () from /lib64/libc.so.6
#2 0x00007f71b5d463aa in _int_realloc () from /lib64/libc.so.6
#3 0x00007f71b5d47d46 in realloc () from /lib64/libc.so.6
#4 0x00007f71bad0dc2a in sdsMakeRoomFor (s=, t=0x7f71943225a0, len=405) at sds.c:97
#5 sdscatlen (s=, t=0x7f71943225a0, len=405) at sds.c:131
#6 0x00007f71bad0b1d0 in __redisAppendCommand (c=0x7f71940038c0, cmd=, len=) at hiredis.c:1186
#7 0x00007f71bad0bef7 in redisvAppendCommand (c=0x7f71940038c0, format=, ap=) at hiredis.c:1206
#8 0x00007f71bad0bfd8 in redisAppendCommand (c=, format=) at hiredis.c:1220

[pietern]

Both of these traces look like they are caused by something in the allocator; tcmalloc in the top one, and the libc allocator in the bottom one. It is more likely that you have bad memory rather than this being a bug in the allocator code. You can check the memory on your machine with memcheck, or with the built-in memory checker in Redis.

Execute it with: redis-server --test-memory.

[goelvivek]

But I got 3 crashes in last 2 days on 3 different machines. It was working
fine earlier from 6 months.
Note: I got crash in redis client hiredis.
The command you have given seems to be for server. Will firing that
command help ?

regards
Vivek Goel

On Tue, Jun 11, 2013 at 9:18 PM, Pieter Noordhuis
[email protected]:

Both of these traces look like they are caused by something in the
allocator; tcmalloc in the top one, and the libc allocator in the bottom
one. It is more likely that you have bad memory rather than this being a
bug in the allocator code. You can check the memory on your machine with
memcheck, or with the built-in memory checker in Redis.

Execute it with: redis-server --test-memory.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/110#issuecomment-19270665
.

[pietern]

You should run that command on the machines where the crash happened, not on the server.

It doesn't look like hiredis itself is the problem here; the crash is caused by the allocator.

[monRedis]

thanks,i have solved it

2013/6/12 Pieter Noordhuis [email protected]

You should run that command on the machines where the crash happened, not
on the server.

It doesn't look like hiredis itself is the problem here; the crash is
caused by the allocator.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/110#issuecomment-19280678
.

[pietern]

How did you solve it in the end? What was the cause for these crashes?

[andoryu-]

I ran into basically the same issue(apart from the fact that glibc is used instead of tcmalloc), which turned out to be a normal heap corruption bug. If the segfault happened randomly at any place using malloc, don't blame hiredis.(Although hiredis uses realloc a lot, which makes it more likely to trigger the problem)

[Denny007]

@monRedis
I have the same issue, would you tell me how to fix it

[sambk]

Hi monRedis and Denny007,
I also have the same issue. Could you share your experience how to solve the problem?

[sambk]

For those who have the same issue, I have fixed my problem. The answer is that the cause is my own memory bug not related to hiredis lib. Perhaps hiredis use many dynamic allocations, the program is prone to crash inside it.