Random processes locking.

[mjm]

One of my systems keeps locking up on random processes. Last night it was Bind, this morning it's Apache, later today it might be MySQL. It's a very fun guessing game of massive downtime for my system....

When a process decides to lock, there's nothing that I can do to make it stop. 'kill -9 XXXX' has no effect. The process has no parent to try and kill. It simply freezes and slowly causes the load on the system to creep higher and higher (although performance doesn't appear to be effected). Once a process has locked, the system won't allow me to SSH in, HTTP doesn't respond, DNS queries go unanswered... If I'm lucky enough to have an open SSH connection when this happens and I try to reboot, the system freezes completely and must be restarted manually (pull the plug!)...

Anyone ever experience anything like this? It smells like a hardware issue to me (because of the randomness)...

 

[vandal]

What is in your /etc/resolv.conf ?

I had something kind of similar using my datacenters DNS servers and then they changed the ip addresses of them and it messed everything up.

 

[nobaloney]

Vandal, the lack of working DNS in resolv.conf should caue timeouts, but not lockups.

mjm, you might want to bring this question up on a forum devoted to your OS. kill -9 should always force kill a process.

Jeff

 

[vandal]

It sure caused lockups and other major problems for me. Exim threads sticking forever, spam assassin completely stopped working and somehow brought httpd down as well. Not sure how or why, but once I finally figured out that they had changed their DNS ips and adjusted to their new ones everything worked perfect.