RPC errors after NFS server restart

[email]Paul.Porcelli@nsc.com[/email] (Paul Porcelli) wrote in message news:<99d5044b.0310140149.327e8f1b@posting.google.com>...[color=blue]
> Hi folks,
> We are running a Linux NFS server (Redhat 7.2).
> Nfsutils version is : 1.0.5-1
>
> Most of our clients are Solaris 8 SUN boxes.
>
> If I stop NFS services on the Linux server (rpc.mountd nfsd rpc.quotad
> rpc.statd lockd) and then restart them soon after we are seeing a
> problem
> with the filemgr application (the OpenWindows file manager).
> When you start the app it takes a long time to start. I then scroll
> down the files in my home area (automounted) and it starts to hang.
> In the console window where I started the app I see RPC timeout
> errors.
> Eventually the filemgr comes back to life with my home area files
> visible
> but then if I scroll down again I see the same problem.
>
> Would be most grateful for any help with this
>
> Many thanks
> Paul Porcelli[/color]

Hi Paul, are you seeing stale nfs handles on your unix clients after
you stop and restart nfs services on the linux server? could be that
someone had a filesystem from your nfs server mounted and had files
open. This would cause the nfs client to think it still had the nfs
filesystem mounted even though it would not be accessable (or even
really nfs mounted at this time).. This would cause a stale nfs
handle on the client machine and could cause rpc timeouts until the
automount point times out.

I don't know your system layouts so it's really hard to troubleshoot
without knowing if your home area files are on the nfs server or nfs
mounted from elsewhere. Also, where are you seeing problems from the
filemgr app? on the linux system or on one of the unix clients? is
the filemgr app possibly nfs mounted from elsewhere or are you using
it locally?

I'm just trying to better understand the problem.

Sharona

RPC errors after NFS server restart - Why ?