Here is the deal, I run a sizable website and do daily backups every night. Ever since I've moved to my new server, I've noticed that after doing backups my load average jumps to 24+ (I have 24 cores), while it is normally 5-6. It seems mysql may be working hard and table corruption occurs (Total fragmented tables: 26). In addition, I believe it is causing a memory dump because (26099120k used, 6856320k free) and at most 40% has been allocated with the running processes listed in TOP.

I believe a few tables are getting corrupted, and causing over load and a memory leak. The database is around 4 gGB (both MyISAM and InnoDB) and I use mysqldump to dump the database. Should I repair the tables and shut down the webserver while backing up?

Operating System: Centos Mysql version: 5.0.77 (I need to upgrade it)

1 Answer 1


Fragmentation is not corruption.

By all means, fix the fragmentation (optimize table table_name; - be warned, this will lock the tables during the operation) - it may be a major contributor to your load while backups are not running, as well as during the backups.

I would seriously recommend running the database on a separate system from your web server, instead of just getting progressively more massive servers on which to run the whole stack. Should make it a lot easier to figure out where the bottlenecks exist in your application, and allow you to allocate resources more specifically (instead of forcing MySQL and your web server to contend for the same RAM).

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .