[SASAG] NFS Timeouts any Tips?
czimmer at wczimmerman.dyndns.org
Thu Jun 5 14:58:39 PDT 2008
How many systems have this mounted as well? I've seen issues where
there were not enough nfsd processes running to serve the requests.
The default in most systems is something like 8 processes, so you can
quickly run out.
On Thu, Jun 5, 2008 at 2:06 PM, Nick Webb <webbn at acm.org> wrote:
> Hi All -
> On a Linux email server (postfix/courier-imap) I'm NFS mounting /home
> and seeing the occasional error like:
> Jun 5 11:37:06 localhost kernel: [ 9924.401601] nfs: server
> 192.168.18.2 not responding, timed out
> At first I suspected network issues, so to isolate that I've put in a
> crossover cable between the two hosts and setup a dedicated private
> network just for NFS. Problem still happens. Next I read that
> perhaps higher rsize and wsize arguments could cause this, especially
> due to bad network equipment, so I went ahead and tried
> rsize/wsize=1024 instead of 4096, no dice.
> Finally I was mounting the shares as "sync,soft" which could be bad
> for email, so I changed that to "sync,hard,intr", but still no dice.
> These errors don't seem major, but I'd like to get them fixed. Users
> occasionally see a error on their mail clients when connecting to
> /etc/exports on the server:
> /home 192.168.18.1(rw,sync,no_root_squash)
> /etc/fstab on the client:
> 192.168.18.2:/home /home nfs
> rw,sync,auto,hard,intr,rsize=1024,wsize=1024,timeo=14 0 0
> Any suggestions? It's been a while since I've done anything major
> over NFS, so I'm sure I'm a bit rusty. If it helps the client server
> is Ubuntu Dapper Server (6.04) and the server is Ubuntu Hardy Server
> (8.04). Neither server get's very loaded...
> Members mailing list
> Members at lists.sasag.org
More information about the Members