Redis crush

[s-tchaikovsky]

Running version 2.8.0-rc1 on Ubunut 12.04

Getting following error while setting about 100 values with MSET command

*** stack smashing detected ***: ./redis-server *:6379 terminated
[2726] 26 Jul 15:00:44.137 #

=== REDIS BUG REPORT START: Cut & paste starting from here ===
[2726] 26 Jul 15:00:44.140 # Redis 2.7.101 crashed by signal: 11
[2726] 26 Jul 15:00:44.141 # Failed assertion: (:0)
[2726] 26 Jul 15:00:44.142 # --- STACK TRACE
Segmentation fault (core dumped)

Tried that on different servers and virtual machine. 100% reproduction

Upd:

Here's link to script i'm trying to execute

https://docs.google.com/file/d/0B6nhA34lmPYJbUlRVjJiWmxHelk/edit?usp=sharing

[antirez]

Thanks for reporting, please could you specify an example of script that triggers this issue?

[s-tchaikovsky]

Here's what i get with GDB:

Program received signal SIGSEGV, Segmentation fault.
0x00987b19 in ?? () from /lib/i386-linux-gnu/libgcc_s.so.1
(gdb) bt
#0 0x00987b19 in ?? () from /lib/i386-linux-gnu/libgcc_s.so.1
#1 0x009886e1 in _Unwind_Backtrace () from /lib/i386-linux-gnu/libgcc_s.so.1
#2 0x00f59007 in backtrace () from /lib/i386-linux-gnu/libc.so.6
#3 0x00ec2310 in ?? () from /lib/i386-linux-gnu/libc.so.6
#4 0x00f58dd5 in __fortify_fail () from /lib/i386-linux-gnu/libc.so.6
#5 0x00f58d8a in __stack_chk_fail () from /lib/i386-linux-gnu/libc.so.6
#6 0x0806f735 in replicationFeedSlaves (slaves=0x69746341, dictid=975335023,
argv=0x736c6166, argc=1143090277) at replication.c:288
#7 0x6e497349 in ?? ()
#8 0x69746341 in ?? ()
#9 0x3a226e6f in ?? ()
#10 0x736c6166 in ?? ()
#11 0x44222c65 in ?? ()
#12 0x6c707369 in ?? ()
#13 0x79547961 in ?? ()
#14 0x3a226570 in ?? ()
#15 0x726f4622 in ?? ()
#16 0x22796e41 in ?? ()
#17 0x7349222c in ?? ()
#18 0x69746341 in ?? ()
#19 0x3a226576 in ?? ()
#20 0x65757274 in ?? ()
#21 0x7349222c in ?? ()
#22 0x776f6853 in ?? ()
#23 0x743a226e in ?? ()
#24 0x2c657572 in ?? ()
#25 0x50734922 in ?? ()
#26 0x69646e65 in ?? ()
#27 0x3a22676e in ?? ()
#28 0x736c6166 in ?? ()
#29 0x0a0d7d65 in ?? ()
#30 0x0d313124 in ?? ()
#31 0x6e61620a in ?? ()
#32 0x5f72656e in ?? ()
#33 0x39303633 in ?? ()
#34 0x00000a0d in ?? ()
#35 0x00000000 in ?? ()

[antirez]

Hello, thanks. What is the way you use the file "ban" attached to your file? At a first glance it does not seem to be properly formatted as as the end it misses the final quote character. Please could you provide a very small program reproducing the issue? Thank you.

[antirez]

Not needed, sorry, I'm now able to reproduce it, I forgot to attach the slave when simulating the large MSET thing. Thanks for reporting.

[antirez]

Fixed, I'll close the issue after regression tests are implemented.

[s-tchaikovsky]

Sorry for the final quote, just forgot it.
It looks wierd, but when i exec mset manually (bash script) with that set of data it doesn't fail, but when i try to do it programmically it crushes redis master node, while replic stays alive. We use ServiceStack.Redis client library.

Here's link to gdb core file:

https://docs.google.com/file/d/0B6nhA34lmPYJUElXT2E5bTMxQ0U/edit?usp=sharing

[antirez]

@s-tchaikovsky thanks, the latest version of Redis 2.8 and unstable are now fixed, however I'll release a proper 2.8 release candidate 2 version in the next days. Closing the issue now that there is a regression test. Thanks for your help.