Weird Server Behaviour

Cantab.

Server keeps crashing

I've a dedicated server that's annoying the hell out of me.

Every 24 hours or so, the server decides to stop serving http and https requests. The email seems to be unaffected thank God.

Usually doing a
sudo apache2ctl restart
sudo /etc/init.d/mysql restart

does the trick

However sometimes I've to do a complete restart.

Anyone have any ideas? This is a really annoying problem for me

feylya

Anything appearing in the logs?

Probably better off posting this in the Server and Systems forum

Ciaran187

Make sure you are closing your DB connections every time. Inspect the SQL process list using

SHOW PROCESSLIST

I'm assuming you're including a footer in every page. Make sure the connection is closed here. Like feylya said, logs would help. Just before it starts rejecting connections.

CreepingDeath

If you're using a database, are you closing all the database connections and statements after each use ?

You could be losing database connections or cursors, until eventually you get connection refused errors.

You could have a table lock issue, if you've two processes attempting to perform a transaction at the same time.

Do you have need to have a static IP address for your server.

Could be anything... server out of memory, database resources, table locks.

Cantab.

CreepingDeath wrote: »

If you're using a database, are you closing all the database connections and statements after each use ?

You could be losing database connections or cursors, until eventually you get connection refused errors.

You could have a table lock issue, if you've two processes attempting to perform a transaction at the same time.

Do you have need to have a static IP address for your server.

Could be anything... server out of memory, database resources, table locks.

Really? Eek...

It's really troubling me. The worry is just that: it could be anything!

I run a backup script at 5 in the morning each day. I ran it again manually this afternoon. Just tried a http request now on the server and it's gone down again! Did an apache restart and a mysql restart and it's ok again.

I'll check to see if it's gone down again tomorrow after this morning's backup. That might help me narrow down the search.

Anyone got any other suggestions? I seriously doubt it's server out of memory. Can you expand on database resource issues and table locks please? I'm running roundcube email and postfix SQL.

Will I just have to re-image the entire server and build her back up again?

She's not doing much at the moment apart from hosting email and serving the odd web page. Just ticking over really.

Cantab.

Found this in my apache2 logs:

[Thu Mar 19 07:05:13 2009] [error] (2)No such file or directory: Cannot create SSLMutex with file `/var/run/apache2/ssl_mutex' Configuration Failed

Evil Phil

Seems fine in here but if you do want me to move it over to Servers and Systems let me know Cantab.

WizZard

Does the directory /var/run/apache2 exist?

Also, can you post the following lines of your ssl.conf

LockFile
AcceptMutex

Does your backup script 'clean up' anything?