Stale NFS file handles on 8.x amd64

[Adam McDougall]

On 11/30/10 02:49, Doug Barton wrote:
> On 11/29/2010 17:06, Adam McDougall wrote:
>> I've been running dovecot 1.1 on FreeBSD 7.x for a while with a bare
>> minimum of NFS problems, but it got worse with 8.x. I have 2-4 servers
>> (usually just 2) accessing mail on a Netapp over NFSv3 via imapd.
>
> There are a whole lot more variables that I haven't seen covered yet.
> Are you using TCP mounts or UDP mounts? Try toggling that setting and
> see if your performance increases. Are you using rpc.lockd, or not? Try
> toggling that. What mount options are you using other than TCP/UDP? What
> does the network topology look like?
>
> It's very likely that we can help you here, but more information is needed.
>
>
> Doug
>

I am using tcp mounts, the mounts in question are either using
rw,bg,tcp,nosuid or rw,bg,tcp,noexec in fstab which is what I was using 
in 7.x except for intr.  I'm much more concerned about the corruption 
rather than the performance at this point but I could try UDP when I get 
a chance, although I wasn't using it on 7.  I am running lockd and statd 
and was on 7 too.  I was using options NFSLOCKD on both
7 and 8.  The Netapp I am accessing is on the same local /24 subnet and 
it does not traverse any firewalls or routers to get there.  Two of the 
four clients are on the same switch as the NFS server, the other two 
clients are on a different layer 2 switch but same vlan.

Today is a bit busy so I may only have time for discussion or simple 
non-invasive changes during the day but I'll compile a list of 
suggestions at least.  Thanks.