Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have a couple of Centos linux servers, that have a very simple task, they run nginx + fastcgi for php , and some NFS mounts between them, readonly

They have some RPC commands to start some downloading processes with wget, nothing fancy , from a main server, but their behavior is very unstable, they simply go down, we tried to monitor ram , processor usage, even network connections, they don't load up so much, max network connections up to... 250 max, 15% processor usage and memory , well, doesn't even fill up, 2.5GB from 8GB max ,

I have no ideea why can a linux server go down like that, they aren't even public servers, no domain names installed no public serving, for sites.

The only thing that I've discovered was that if i didn't restart the network service every couple of hours or so... the servers were becoming very slow, starting apps very slow, but not repoting a high usage of resources...Maybe Centos doesn't free the timeout connections, or something like that...It's based on Red Hat right?

I'm not a linux expert , but I'm sure that there are a few guys out there that can easily have an answer to this , or even have some leads to what i can do ...

I haven't installed snort, or other things to view if we have some DOS attacks, still the scheduled script that restarts the network each hour should put the system back online, and it doesn't....

Thank you in advance

share|improve this question

migrated from stackoverflow.com Mar 28 '10 at 22:10

This question came from our site for professional and enthusiast programmers.

    
Which version of CentOS? Are you fully updated? What type of hardware? –  Dave Mar 29 '10 at 3:33

3 Answers 3

old kernel, ext3 bug when it hits an extent. Filesystem is likely corrupt somewhere that is getting hit once in a while, possibly when weblogs are written.

upgrade your ext3tools to the latest version, fsck the filesystem. See if there is a newer kernel release than the one you are running. While ext3 is stable, there have been some very recent patches to ext3 (and xfs) that you might be hitting.

share|improve this answer

I've sent the tech to the site, and this is what the server shows on the screen when the server was down (see image link)

http://ft.beejive.com/icq/648/480731375/img0328152705_kv044e.jpg

Which is very wierd, no kernel panic in messages, and i don't know what could it be

share|improve this answer
    
I'm considering onto installing Gentoo, because it compiles itself onto the machine, maybe it handles better the ethernet packages. I understood that the best linux for networking is or was slackware. And for best serving purposes, is freebsd, but still i want a linux not a bsd.. –  PartySoft Mar 29 '10 at 0:32
    
CentOS is based on Red Hat Enterprise Linux. Companies of all sizes use it successfully in production environments. I'm a huge Slackware fan, my favorite distribution to this day. Nevertheless, don't be so quick to blame CentOS. –  Warner Mar 29 '10 at 13:14
    
I'm not judging CentOS, i'm just trying to find a solution to my problem, never the less Slack is one of the most stable when it comes to networking..or so was in the past when i used it, –  PartySoft Apr 5 '10 at 18:09

Jsut a random thought that you might look into your DNS entry in the /etc/resolv.conf file and other networking places in the filesystem.Try to "dig" your nameserver and well as the router get the information out of that.

As you said I would love to look into the /etc/sysconfig/network file and related places to hunt down.

Cheers!

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.