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'm having something that I can not understand. I quite new on lustre.

n4 and n5 are two I/O servers, each one taking care of different datasets.

When one of them goes down, the other is supposed to handle the service as backup node. But n4 goes down without having heavy loads, and when n5 gets the service its cpu and memory goes to the top and performance goes to the bottom.

Can someone tell me which lines to look at , and from here I can start checking this issue?

Here the a 30 minutes syslog with lustre related messages from n4 : http://pastebin.com/q1iGwDxw

Here almost the same 30 minutes from n5. http://pastebin.com/4Bg5repa

Many thanks.


share|improve this question
I am not a lustre user but I see this "LDISKFS-fs warning: maximal mount count reached, running e2fsck is recommended" in the logs about a hundred times. Have you tried that? Is it possible that there is corruption on your volumes causing lustre grief? –  polynomial Aug 25 '11 at 17:22
I have been told that e2fsck is quite dangerous. Have you done it before on a lustre fs? –  marc.riera Aug 26 '11 at 9:48
Well like I said I'm not a lustre user but even on their site they indicate this may be necessary: wiki.lustre.org/index.php?title=Fsck_Support I really think you should research that and fix that error message before looking for more complex solutions to what is probably a simple issue of a corrupt FS causing problems. –  polynomial Aug 26 '11 at 14:03

Your Answer


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

Browse other questions tagged or ask your own question.