MySQL Bugs: #13962: InnoDB makes MySQLd crash after 5 minutes of operation

Bug #13962	InnoDB makes MySQLd crash after 5 minutes of operation
Submitted:	12 Oct 2005 15:29	Modified:	12 Nov 2005 16:36
Reporter:	[ name withheld ]	Email Updates:
Status:	No Feedback	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:	3.23.58	OS:	Linux (Fedora core 2)
Assigned to:		CPU Architecture:	Any

Description:
After (re-)starting MySQLd with InnoDB enabled, MySQLd crashes after a few minutes.

After starting MySQLd the /var/log/mysqld.log shows:

051012 16:56:25  mysqld started
051012 16:56:26  InnoDB: Database was not shut down normally.
InnoDB: Starting recovery from log files...
InnoDB: Starting log scan based on checkpoint at
InnoDB: log sequence number 0 2297501
InnoDB: Doing recovery: scanned up to log sequence number 0 2297891
051012 16:56:26  InnoDB: Starting an apply batch of log records to the database...
InnoDB: Progress in percents: 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 
InnoDB: Apply batch completed
051012 16:56:27  InnoDB: Flushing modified pages from the buffer pool...
051012 16:56:27  InnoDB: Started
/usr/libexec/mysqld: ready for connections

Then after a few minutes none of my PHP-sites using MySQL works anymore and at that moment the command "service mysqld status" returns:
"mysqld dead but subsys locked"

The logfile appears to be expanded by:
Number of processes running now: 0
051012 16:59:34  mysqld restarted
051012 16:59:35  InnoDB: Database was not shut down normally.
InnoDB: Starting recovery from log files...
InnoDB: Starting log scan based on checkpoint at
InnoDB: log sequence number 0 2298018
InnoDB: Doing recovery: scanned up to log sequence number 0 2298018
051012 16:59:35  InnoDB: Flushing modified pages from the buffer pool...
051012 16:59:36  InnoDB: Started
/usr/libexec/mysqld: ready for connections

Number of processes running now: 0
051012 17:04:32  mysqld restarted
051012 17:04:32  InnoDB: Database was not shut down normally.
InnoDB: Starting recovery from log files...
InnoDB: Starting log scan based on checkpoint at
InnoDB: log sequence number 0 2302379
InnoDB: Doing recovery: scanned up to log sequence number 0 2302379
051012 17:04:33  InnoDB: Flushing modified pages from the buffer pool...
051012 17:04:33  InnoDB: Started
/usr/libexec/mysqld: ready for connections

Number of processes running now: 0
051012 17:15:15  mysqld restarted
051012 17:15:25  InnoDB: Database was not shut down normally.
InnoDB: Starting recovery from log files...
InnoDB: Starting log scan based on checkpoint at
InnoDB: log sequence number 0 2302771
InnoDB: Doing recovery: scanned up to log sequence number 0 2302771
051012 17:15:27  mysqld ended

My (Plesk) Control Panel uses InnoDB-tables, so I don't know how many queries have been executed on InnoDB tables at that moment. It appears that MySQLd has restarted a couple of times and after that it ended itself.

When I comment all my innodb-lines in /etc/my.cnf MySQLd runs perfectly well, but Plesk doesn't work, which isn't the ideal situation either.

All this started a week ago now. I haven't done anything "strange" (installing, de-installing, configuring, etc.) before then, so I don't know how to repeat this.
I did edit my /etc/my.cnf since the first "crash", but it didn't help.

Before the first crash, the my.cnf was:

[mysqld]
safe-show-database
innodb_data_file_path=ibdata1:10M:autoextend
datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock

[mysql.server]
user=mysql
basedir=/var/lib

[safe_mysqld]
err-log=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid

and now it is:

[mysqld]
safe-show-database
innodb_data_file_path=ibdata1:10M:autoextend:max:100M
datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock
;skip-innodb
set-variable = innodb_additional_mem_pool_size=10M
;set-variable = innodb_buffer_pool_size=50M
;set-variable = innodb_log_file_size=5M
;set-variable = innodb_log_buffer_size=3M
;innodb_flush_log_at_trx_commit=1

[mysql.server]
user=mysql
basedir=/var/lib

[safe_mysqld]
err-log=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid

I have tried mysqld running with uncommenting all combinations of the commented lines above, but nothing worked out.

How to repeat:
I have no idea how to repeat this.

Suggested fix:
I have no idea how to fix this.

Thank you for a problem report.

So, InnoDB fails started a week ago? Would you, please, describe what was performed at that time and post the error log for that period, not only for today? You may compress it (if it is large) and post using the File tab. Or just post a part starting from the first problems with InnoDB reported.

MySQLd.log.1 (as old as I could find)

Attachment: mysqld.log.1 (text/plain), 54.42 KiB.

mysqld.log (newest)

Attachment: mysqld.log (application/octet-stream, text), 11.37 KiB.

Hi!

I do not see anything in the .err log that would indicate that the problem has to do with InnoDB. mysqld may be shutting down for some other reason.

Are you running on Linux/x86 or Linux/AMD64? On Linux/x86 mysqld should print a stack trace if there is a seg fault.

Do you have enough RAM in your computer?

If you run CHECK TABLE on your tables do they appear ok?

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".