Bug #40898 maria: Server crashes in sysbench OLTP_RW test at ma_state.c:94
Submitted: 20 Nov 2008 17:37 Modified: 10 Dec 2008 21:16
Reporter: Alexey Stroganov Email Updates:
Status: Can't repeat Impact on me:
None 
Category:MySQL Server: Maria storage engine Severity:S1 (Critical)
Version:6.0.8-release OS:Any
Assigned to: Michael Widenius CPU Architecture:Any

[20 Nov 2008 17:37] Alexey Stroganov
Description:
The crash with following backtrace was observed in OLTP_RW sysbench test:

Program terminated with signal 11, Segmentation fault.
#0  0x00002b5078fa24c5 in pthread_kill () from /lib64/libpthread.so.0
#0  0x00002b5078fa24c5 in pthread_kill () from /lib64/libpthread.so.0
#1  0x000000000064944e in handle_segfault (sig=11) at mysqld.cc:2673
#2  <signal handler called>
#3  _ma_setup_live_state (info=0x1b9c2f0) at ma_state.c:94
#4  0x0000000000a23e1d in ha_maria::external_lock (this=0x1b65428, thd=0x1a68ef0, lock_type=0) at ha_maria.cc:2205
#5  0x000000000073c1aa in handler::ha_external_lock (this=0x1ada350, thd=0x1ada350, lock_type=0) at handler.cc:5298
#6  0x00000000006422ab in mysql_lock_tables (thd=0x1a68ef0, tables=<value optimized out>, count=<value optimized out>, flags=
<value optimized out>, need_reopen=0x4007fe27) at lock.cc:411
#7  0x000000000068d832 in lock_tables (thd=0x1a68ef0, tables=0x0, count=<value optimized out>, flags=0, need_reopen=0x4007fe2
7) at sql_base.cc:4465
#8  0x0000000000694295 in open_and_lock_tables_derived (thd=0x1a68ef0, tables=0x1af9d28, derived=true, flags=0) at sql_base.c
c:4166
#9  0x0000000000656b77 in execute_sqlcom_select (thd=0x1a68ef0, all_tables=0x1af9d28) at mysql_priv.h:1620
#10 0x000000000065a9de in mysql_execute_command (thd=0x1a68ef0) at sql_parse.cc:2066
#11 0x00000000006d4029 in Prepared_statement::execute (this=0x1af8810, expanded_query=<value optimized out>, open_cursor=fals
e) at sql_prepare.cc:3578
#12 0x00000000006d6ffc in Prepared_statement::execute_loop (this=0x1af8810, expanded_query=0x40081470, open_cursor=false, pac
ket=<value optimized out>, packet_end=<value optimized out>) at sql_prepare.cc:3245
#13 0x00000000006d76a5 in mysql_stmt_execute (thd=0x1a68ef0, packet_arg=0x1abad71 "\001", packet_length=157514) at sql_prepar
e.cc:2467
#14 0x000000000065febe in dispatch_command (command=COM_STMT_EXECUTE, thd=0x1a68ef0, packet=0x1abad71 "\001", packet_length=1
5) at sql_parse.cc:960
#15 0x0000000000652a72 in handle_one_connection (arg=<value optimized out>) at sql_connect.cc:1156
#16 0x00002b5078f9e193 in start_thread () from /lib64/libpthread.so.0
#17 0x00002b507981645d in clone () from /lib64/libc.so.6
#18 0x0000000000000000 in ?? ()

How to repeat:
Instructions and test case will be provided later.
[9 Dec 2008 15:58] Guilhem Bichot
looks similar to what Monty has just fixed, so assigning to him.
[10 Dec 2008 21:16] Michael Widenius
Repeated the test run on MySQL-5.1-Maria with the fix for history_state without any problems.

I assume this is now fixed in MySQL-5.1-Maria; It should be in MySQL-6.0 shortly