MySQL Bugs: #61453: Segmentation Fault

Bug #61453	Segmentation Fault
Submitted:	8 Jun 2011 21:35	Modified:	18 Jun 2011 11:02
Reporter:	Chris Wagner	Email Updates:
Status:	Duplicate	Impact on me:	None
Category:	MySQL Server: General	Severity:	S2 (Serious)
Version:	5.5.13	OS:	Linux (Ubuntu)
Assigned to:		CPU Architecture:	Any
Tags:	myisam_use_mmap, mysqld got signal 11, segfault, segv, Signal 11, SIGSEGV

Description:
MySQL has begun crashing on us.  Basically our application is Perl driven and gathers stats on a large database of data.  Matching data is collected from source MyISAM tables into a temp table.  Stats are generated from that and saved to InnoDB reporting tables.  There is no one thing that causes the crash.  Scripts will be at random points in the process and the server just dies with "mysqld got signal 11".  The crash frequency is daily when a job is running.  I will attach the error log output.

Our box is a Dell 8-way Xeon (hyperthreading) with 64 GB of RAM and hardware RAID going to 15k drives.

OS info:
Linux version 2.6.35-25-server (buildd@allspice) (gcc version 4.4.3 (Ubuntu 4.4.
3-4ubuntu5) ) #44~lucid1-Ubuntu SMP Tue Jan 25 19:34:09 UTC 2011
Ubuntu 2.6.35-25.44~lucid1-server 2.6.35.10

Thanks.

How to repeat:
Run queries.

Error log excerpts

Attachment: crashes.txt (text/plain), 7.03 KiB.

my.cnf

Attachment: config1.txt (text/plain), 1.18 KiB.

Thank you for the bug report. Could you please elaborate what query provoked the crash (a dump of the table(s) and query could help to repeat it on our side), your my.cnf and the back trace. Thanks in advance.

Please disregard some question I did before I was editing simultaneously with you. However the dump of table(s) would be nice. Thanks.

Thanks for responding.  Each crash occured while accessing a different table so I'm not sure what I would show u.  Once it was on SELECT INTO OUTFILE from MyISAM.  Another time inside an InnoDB transaction.  The MyISAM tables are 8x8 partitioned due to size.  The InnoDB tables are pretty vanilla standalone tables.

Just experienced a new crash.  This time however it printed a message to the console about an invalid pointer.  *see attachment*

console messages and error log for crash

Attachment: crashes2.txt (text/plain), 20.10 KiB.

Please, try to comment out:

myisam_use_mmap

in my.cnf, restart and check if you'll ever get any similar crashes.

Did that.  Now waiting to see how it acts.

Please re-open when you are done with the result of myisam_use_mmap. Thanks.

It's been several days and I've run several large jobs and there hasn't been any new crashes.  It would seem that myisam_use_mmap is indeed the source of the crashing.  I don't recall anything in the manual about that being a stability risk.

see bug #48726 ...

Duplicate of bug #48726.