2013/2/1 <[email protected]>
> Hi > > Reading this mroning the other opened thread about "Problem with exportfs" > , I'm thinking about another problem I faced with exportfs : > > on rhel, when command exportfs -u successfully returns, you may think you > could umount the FS exported, but this is not always true, even if fuser > or lsof does not return any process using the FS, sometimes it can't be > unmounted > just after the exportfs -u . That leads to a very annoying problem with > Pacemaker, when you have for example a group with in order : > File system resource > exportfs resource > when stopping the group, sometimes (but often) the Stop op on Filesystem > fails, hence Pacemaker fences the node (on-fail=fence) > > Just a precision : I 've only NFS v3 clients, none v4 (meaning I think > that le leastime for nfsv4 should not exist in that case) > > I know that it seems to be more a problem of exportfs -u command, than a > problem on exportfs ocf script but ... > > In deed the /proc/fs/nfsd/nfsv4leasetime parameter have its role in v4 (in terms of stop-timeout) but in v3, maybe some other mechanism ? It might be useful to force /bin/systemctl restart nfs-server.service (or whatever LSB service) At each exportfs command, for me it does discharge the directory that can then be unmounted most of the time. Maybe there's a clever way to do it. _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
