From: Tim Cutts (tjrc@sanger.ac.uk)
Date: Tue Aug 08 2006 - 04:21:55 EDT
Thanks to Eric Werme, who agreed that my suggestion was probably
safest, probably combined with /sbin/init.d/nfs cluster_stop, but
also thought it was reasonably safe to just delete the file anyway.
As it happened, a network problem caused the cluster to crash about
an hour after I sent my message, so a golden opportunity to clean up
the file presented itself anyway!
Tim
On 4 Aug 2006, at 12:44 pm, Tim Cutts wrote:
> Is there a procedure for purging old entries from /etc/mountdtab on
> TruCluster NFS servers?
>
> The file has over 23,000 entries on one of our NFS servers,
> including entries for several hundred Linux clients which no longer
> exist.
> This causes the RPC mountlist linked list to be so large that some
> clients (notably am-utils on Linux) run out of stack space when
> trying to work their way through it.
>
> Is there a safe way to purge entries from this file? Is it
> sufficient to:
>
> 1) /sbin/init.d/nfs stop
> on all TruCluster nodes
> 2) Remove all the obsolete entries
> 3) /sbin/init.d/nfs start
> on all nodes
>
> ?
>
> Thanks in advance.
>
> Tim
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:50:31 EDT