[DRBD-user] Stale NFS file handle vs. NFS-Server-README.txt

Eugene Crosser crosser at rol.ru
Wed Jun 9 14:52:25 CEST 2004

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


On Wed, 2004-06-09 at 16:36, Peter Kruse wrote:

> > - do *not* ever unexport the filesystem when you stop NFS server
> > - keep /var/lib/nfs in the exported filesystem (e.g. symlink there)
> 
> I've seen this suggestion several times, and I don't know who came
> up with this ... idea.  What are going to do in an active/active
> Cluster?  What if one node fails over to the other?  Better is to
> keep /var/lib/nfs where it is and merge information to the
> rmtab file.  This is how we do it anyways.

You just do it in a suboptimal way.  When the cluster switchover takes
place, NFS clients get IO error or even need to remount the filesystem. 
When done "properly", switchover is completely transparent to the
clients.  That is, even when switchover takes place when the client is
in the middle of writing a big file, operation will be completed on the
new node and integrity of the content will be preserved.

Speaking specifically about /var/lib/nfs location, the recommendation to
keep it on shared media is only relevant when the filesystem is exported
to network range(s) (as opposed to individual hosts).

Eugene
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20040609/b636ce12/attachment.pgp>


More information about the drbd-user mailing list