MySQL Signal 10 on FreeBSD 6.0-RELEASE-P6

[Erik Kristensen]

Thanks for the reply. I apologize, I should have been more clear with my 
first paragraph in my email. I have found several other somewhat similar 
problems with MySQL on FreeBSD 6.0 but with the MySQL 4.1.x series.

Nothing in the Bug report you suggested bears much resemblance at first 
glance. I originally thought this was a too many connections for the 
server (physical server) problem, so I dropped the max connections to 20 
and still got a signal 10. I also played with other numbers as the max 
connections but no matter what about once an hour a SIGBUS error would 
appear.

I have enabled --log and --core-file on my mysql server and I am trying 
to determine if a specific query is causing the problem, however I 
highly doubt this. Hopefully when it crashes the core file will be able 
to shed some light on the problem at hand.

Just as I was typing I had tail running on the log file and error log 
files and I caught a signal 10, at the time there was no query that 
stood out that hadn't been queried previously. If you like I could send 
the queries that preceded the signal 10, but I doubt they will do any 
good. There are 4 queries and about 25 quits, in the minute leading up 
to the crash.

I do have a core dump file and I am attempting to gather information, 
however my knowledge on inspecting core dumps is limited, I am willing 
to host the file somewhere if you care to look at it to help determine 
the cause.

Many thanks.

Regards,
-Erik

Greg 'groggy' Lehey wrote:
> On Monday, 31 July 2006 at 11:35:04 -0400, Erik Kristensen wrote:
>> I have an interesting issue with FreeBSD 6.0 and MySQL 5.0. Part of
>> my problem has been discussed serveral times in the past on other
>> mailing lists and it seems there has been fixes for older versions of
>> MySQL.
> 
> From what you say here, there's no reason to believe that your problem
> has been mentioned at all.
> 
>> The MySQL server is currently receiving a SIGNAL 10 about once an
>> hour, which is causing many problems with the innodb databases that
>> we have running on this server.
> 
> Signal 10 is SIGBUS (bus error), an error detected by the hardware.
> The most frequent causes are bugs in the hardware and bugs in the
> software.  The only way to find out what causes the problem is to
> investigate further.  Never make the assumption that any two SIGBUS
> problems are related.  Still, SIGBUS is interesting, because just
> about every such error in FreeBSD is reported as a SIGSEGV (signal
> 11).
> 
>> I am running FreeBSD 6.0-RELEASE-p6 and MySQL 5.0.22.
>>
>> In all the lists I have read regarding FreeBSD, MySQL and signal 10,
>> it all seems to happen with version 6.0 and the 4.1.x series of
>> MySQL.
> 
> I can't confirm that at all.  I've been chasing a number of very
> specific problems for some time, and as far as I can tell, there are
> very few cases where the server crashes.  One case that I'm
> investigating is BUG#12251 (http://bugs.mysql.com/bug.php?id=12251).
> If any of this looks familiar to you, please let me know.
> 
>> Any help in this area would be greatly appreciate as this is our
>> production server having issues. We are a small company with limited
>> resources.
> 
> A start would be the documented method to track down crashes:
> http://dev.mysql.com/doc/refman/5.0/en/debugging-server.html gives
> details.  In particular, the contents of the error log will give you
> some idea, though probably you'll need a debugger to get a usable
> backtrace.
> 
> Greg
> --
> When replying to this message, please copy the original recipients.
> If you don't, I may ignore the reply or reply to the original recipients.
> For more information, see http://www.lemis.com/questions.html
> See complete headers for address and phone numbers.