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.

We're running into a very odd issue w/ a CentOS 6 server.

We have several ifcfg-range scripts setup to add /24's which are tunneled to the server using GRE. These same scripts and tunnels work fine on other server, this is the only one having the problem.

What we found is that as the server starts adding the IPs in the scripts it runs extremely slow, taking 5-10 minutes per IP. As you can imagine after a while the server just locks up completely.

(We ended up having to use a system rescue disk to go in and remove all the range scripts to restore service).

There are no errors in the range scripts and even if there were the default behavior is to send an alert to screen and log and continue loading.

I've never seen a server behave this way, just wondering if there is an OS setting or if this indicates some issue w/ the network hardware?

share|improve this question
    
cat /proc/interrupts when running script; does it indicate high interrupts? Does htop show a lot of system time? If you see high kernel time it may be issue with nic or nic driver. –  josten Apr 16 '13 at 5:48
    
messages log does show a lot of interrupt based activity going on while the server was trying to bind the IPs from the scripts. (To note, it was able to load the IPs but was doing it so slowly it basically locked the server up after a while). We found out we're on a brand new Supermicro Blade server, something they did not advertise when we chose the hardware on the datacenter website. –  linux911 Apr 16 '13 at 6:12

Your Answer

 
discard

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

Browse other questions tagged or ask your own question.