Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I'm seeing a MySQL server running normally, but once it hits high load, it dies without warning -- instantly crashes. It is supporting both a forum and a chat server. It runs both InnoDB and MyISAM tables. The chat server is the issue, I think.

I am having a hard time determining how to fix the problem, and was wondering what monitoring tools are available for MySQL that would let me specifically determine the cause of the crash (that way I can tweak my.ini settings and hopefully make it go away).

I can't find anything conclusively useful in the appropriate .err file.

I'm seeing Application Event Log 100 messages like this:

Event Type: Information

Event Source: MySQL

Event Category: None

Event ID: 100

Date: 11/13/2009

Time: 3:57:52 PM

User: N/A

Computer: HOST

Description: C:\Program Files (x86)\Parallels\Plesk\Databases\MySQL\bin\mysqld-nt.exe: ready for connections. Version: '5.0.45-community-nt-log' socket: '' port: 3306 MySQL Community Edition (GPL)

Will no other logged information.

A current show status like '%thread%'; shows this for output, if that helps:

Delayed_insert_threads 0 Slow_launch_threads 0 Threads_cached 1 Threads_connected 7 Threads_created 53033 Threads_running 1

Since it is running as a 32-bit instance on this 64-bit box, perhaps the MySQL instance is simply running out of memory? I'm sure a 64-bit upgrade would help TREMENDOUSLY...but I need to be sure it will fix the problem.

Any help is greatly appreciated, and if you have any questions, ask away!

share|improve this question
    
+1 good question, upvoted –  Zak Nov 19 '09 at 22:59

1 Answer 1

up vote 0 down vote accepted

I'm not trying to be trite, but you state this: Server hits high load, dies.. My first solution would be (assuming high load is really due to high usage), add another read server in slave mode.

Second, a particular query could be causing the crash. Maybe if you turn on query loggging you can see if the last query before death is consistently the same. Note, this will increase server load in and of itself, and you need to be on top of log rotation so you don't fill your disks.

Also, if you get a slave up, if your primary dies, you can always promote the slave, so it is good redundancy.

share|improve this answer
    
Not taken as trite -- truly. extremely useful –  asteroid Nov 16 '09 at 19:32

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.