# [EMAIL PROTECTED] / 2003-06-03 22:59:47 +0300:
> you can use SHOW INNODB STATUS to look what active transactions you have and
> how many lock structs they have.
>
> Also innodb_lock_monitor helps.
Hi Heikki,
thanks for the reply. I have dropped the database, recreated it,
stopped the server and started it again, turned on the
innodb_lock_monitor, and had a hard time getting it to deadlock.
After some time, it finaly issued the LOCK wait timeout exceeded
error.
I have the error log with innodb_lock_monitor messages, and the
query log. Both were empty when I started the server, and I shut
down the server right after I got the error (looks like the monitor
writes into the error log even when the server doesn't process any
queries). Plus, I was the only user accessing the server.
Unfortunately, I've forgotten to run SHOW INNODB STATUS, is that
info crucial, or can you get it from the monitor? I hope the latter
is the case since the error is intermittent, and sometimes it's
quite hard to make it show up.
I've put the log files online at http://roman.bellavista.biz/mysql/,
since they're quite big.
An interesting fact: I was trying to get it to spit out the error,
and couldn't, then left the server alone for ~30 minutes, and then
got the error immediately: grep the logs for "13:53".
What I don't understand is the TABLE LOCK lines: all the tables are
InnoDB tables, and I thought InnoDB doesn't use table locks. Or does
it. The only SQL statements in the query log that have to do with
locking are two pairs of GET_LOCK()/RELEASE_LOCK() wrapping the
creation of sequence tables.
I will be very grateful for any info.
--
If you cc me or remove the list(s) completely I'll most likely ignore
your message. see http://www.eyrie.org./~eagle/faqs/questions.html
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
