Portal Home > Knowledgebase > Articles Database > strange mysql crash
strange mysql crash
| Posted by MikeQ, 07-06-2009, 02:21 AM |
| I just moved to this server 3 days ago, installed CPanel ect... and noticed mysql randomly crashes...
did some google searches but came up with nothing recent.
I'm running on 4GB ram with:
|
| Posted by relichost, 07-06-2009, 02:43 AM |
| Hi
Anything in the log before
090706 2:03:15 - mysqld got signal 11 ;
Also how did you migrate to this server ? Copy files manually, or use cpanel move ?
Thanks
|
| Posted by ZenMonk, 07-06-2009, 03:44 AM |
| Looks like your innodb tables have been corrupted and has to be fixed. You can start your mysql server using the skip-innodb option in your my.cnf.
|
| Posted by Steven, 07-06-2009, 05:46 PM |
| What is with all the horrible advice on this board?
If he has innodb databases which he very well could because they are becoming more common, they will no longer operate with skip-innodb causing downtime to his sites.
|
| Posted by Aun Muhammad, 07-06-2009, 06:03 PM |
| You can find out further information regarding InnoDB here.
http://linuxchunks.ittweaks.com/2009...pros-and-cons/
|
| Posted by RBBOT, 07-06-2009, 07:11 PM |
| Come on Steven, you've been here long enough to know that if you write a post that contains the word innodb, someone will reply telling you how to turn it off whether you want it off or not. Them's the rules.
As for the actual problem - this may well be a mysql bug - you just have to look at the number of crash-bugs fixed in each point release to see how many there are out there. If it's repeating often, post on the mysql.com forums asking for help.
|
| Posted by Jonathan Kinney, 07-06-2009, 08:04 PM |
| I have seen some worpress related queries trigger a crash like this before, so it would be good to start some logging, or even do an strace on the process so you can get an idea of what happens before it crashes. You can turn on logging for a short period of time and try and track this down too. If it is worpress, or a worpress plugin, or anything else like that, you can probably modify the code to not trigger the bug anymore. I would also suggest submitting bugs to MySQL and the code maker of what ever is causing the query that triggers the crash. In the end, updating the MySQL server or the code may solve the problem.
|
| Posted by MikeQ, 07-06-2009, 10:46 PM |
| What would be the best way to track down the query that could be causing the crashing?
Just got this:
I've got 1.5GB RAM Free as well...
|
| Posted by RHS-Chris, 07-06-2009, 11:17 PM |
| I've been having the same issues as well on one of my servers, for the past couple of days, CentOS with cPanel. Mine seems to be stable right now though. I had a client who's database was corrupted, so, I restored a backup. I've also added in wait_timeout variable into my.cnf, originally set at 10, now at 20.
|
| Posted by Winstyn, 07-07-2009, 12:21 AM |
| This might sound obvious, but are you running the latest version of all the software on your server. By doing a yum -y udpate and also the latest version of mysql.
Upgrading MySQL might help. It looks like a simple Unique selection went bad. Its possible that the table files got garbled during the transfer especially if it was an rsync. Just losing one packet would cause this to happen.
|
| Posted by MikeQ, 07-07-2009, 12:26 AM |
| Can I upgrade MySQL without screwing up CPanel? I know CPanel has this thing wanting to control your server and services...
As for the databases, would doing a mass repair all or something fix that issue?
|
| Posted by Winstyn, 07-07-2009, 12:30 AM |
| This is the query in question here.
So maybe just repairing that table would help. Its always better to make a targeted attack rather than carpet bombing.
That query almost looks like an internal cPanel query.
|
| Posted by Steven, 07-07-2009, 12:34 AM |
| Looks like a whmcs query to me.
|
| Posted by Winstyn, 07-07-2009, 12:42 AM |
| Thanks for that. I wasnt sure of the software and havent ran into enough WHMCS to know the DB schema.
MikeQ if you know your the only one of the server with a WHMCS install just repair all the tables in your WHMCS folder and keep an eye on MySQL that might fix the problem.
If it has problems repairing the tables you will most definitely know thats where the problem is. I would assume you could stand to empty that table if it doesnt repair. After that you can repair it and it should work.
Hope you get it solved!
|
| Posted by koithara, 07-07-2009, 08:48 AM |
| why don't you try some tuning on db part. also enable mysql logging for normal as well as slow queries and watch out for its results. It may help in some way in finding the root cause.
As pointed out by chris, since migration took place very recently, there is a chance for db corruption as well. In WHM under mysql options there is a option to repair databases. try repairing all databases as well. It may help.
|
Add to Favourites 
Print this Article
Also Read
