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 work for a small software house.

We share our developers' home directories over NFS. The primary use for this is to host the users' git repository.

Since last week, there has been a signficant degradation in performance. For example, cmake invocation (used by our build system) has slowed down markedly. I don't have numbers for a build on nfs share from before the slowdown, but the comparison between local disk and nfs for this task is:

Local: 29.3s
NFS: 9m 46.9s (586s)

So we have this job running (rather neatly) 20x slower on the NFS share. Load average on the NFS server is not high. If anything, we have fewer staff working due to people taking annual leave in August. I think it must be either network or server specific because the problem is seen from multiple clients.

What might cause this kind of slowdown?

Edit: Server is Centos 6.3 running on a DL180G6. Kernel is 2.6.32-279.el6.x86_64. Clients are generally Centos 6.2 and 6.3. Network is 1G ethernet with utp cabling. Network mount points are shared via NIS maps. The server is running autofs-5.0.5-54.el6.x86_64.

The export line in /etc/exports on the NFS server is:

/export/home 10.0.1.0/24(rw,nohide,no_root_squash,sync) 10.1.0.0/24(rw,nohide,no_root_squash,sync)

The timestamp on this file does not suggest that this has been edited recently.

share|improve this question

closed as unclear what you're asking by Chopper3, mdpc, Ward, dawud, Dave M Aug 23 '13 at 14:18

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question.If this question can be reworded to fit the rules in the help center, please edit the question.

    
What are your NFS mount options? –  Janne Pikkarainen Aug 20 '13 at 9:07
1  
When you typed out this question did no part of your conscious mind not consider adding some of the pertinent details here? say make/model of whatever your NFS servers is, OS version, config, filesystem, network design - you know, some form of actual detail that might give us a 1% chance of helping you here. This is the technical version of "My car had broken down, it's a blue one". –  Chopper3 Aug 20 '13 at 9:15
    
Harsh...but fair! Server is Centos 6.3 running on a DL180G6. Kernel is 2.6.32-279.el6.x86_64. Clients are generally Centos 6.2 and 6.3. Network is 1G ethernet with utp cabling. Network mount points are shared via NIS maps. The server is running autofs: autofs-5.0.5-54.el6.x86_64 I'll put the share options into a separate comment, because I'm running out of characters. –  user178883 Aug 20 '13 at 9:54
    
The export line in /etc/exports on the NFS server is: /export/home 10.0.1.0/24(rw,nohide,no_root_squash,sync) 10.1.0.0/24(rw,nohide,no_root_squash,sync) The timestamp on this file does not suggest that this has been edited recently. –  user178883 Aug 20 '13 at 9:55

1 Answer 1

The answer turned out to be an error on the RAID card.

Sorry for the inconvenience. I write because of the slim possibilty that someone else might hit something similar.

To recap: build duration slowed down by a factor of ~3, and the duration was consistent +/-10%, hence suggesting that load wasn't the primary factor.

share|improve this answer

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