We've been having some problems with the server that MySQL was running on at work. First, we were running kernel 2.2.14 and MySQL 3.23.33, and kernel 2.2.14 has an I/O error that mysql doesn't like. We upgraded to kernel 2.2.20, and MySQL 3.23.44 in the same day. Yesterday, I discovered that the filesystem was corrupted, so we did fsck'd and restarted and there haven't been any more errors in linux's logs.
Unfortunately, I had another one of my tables go corrupt and lose a row today. The table is in a database that one person uses through a PHP frontend. The funny thing is that I know for a fact that it went corrupt between 8 am and 11 am this morning... and there aren't any queries to the table, -or errors-, in the binary log! (There isn't anything in the errors log, either...)
The table only has 1300 rows in it. I always figured that MySQL is more stable than this. Is there anything I've overlooked, or other reasons why corruption could be happening randomly like this? What kind of tests can I do to troubleshoot further?
Unfortunately, I had another one of my tables go corrupt and lose a row today. The table is in a database that one person uses through a PHP frontend. The funny thing is that I know for a fact that it went corrupt between 8 am and 11 am this morning... and there aren't any queries to the table, -or errors-, in the binary log! (There isn't anything in the errors log, either...)
The table only has 1300 rows in it. I always figured that MySQL is more stable than this. Is there anything I've overlooked, or other reasons why corruption could be happening randomly like this? What kind of tests can I do to troubleshoot further?