[ClusterLabs] Handling NFS lock recovery

[Michael Schwartzkopff]

Am Montag, 20. April 2015, 07:13:52 schrieb Daniel O'Brien:
> I'm trying to figure out the best way to handle NFS lock recovery on an
> Active-Active setup. I've seen some guides suggest lsp and statd but I've
> also found an ra called nfsserver.
> 
> Assuming the need to handle NFSv3 is there a preference between each? Does
> using only NFSv4 make things simpler?

In NFSv4 the client is responsible for claiming and renewing the locks. So if 
the client does not reclaim the lock in the leasetime, the server frees it. 
when a server reboot occures the server waits for the gracetime untils it 
accepts locks. For details see:

https://access.redhat.com/solutions/42868

This setup makes live much easier, since no file sync or other is necessary. 
This only works ins active/passive setups!

If you really need am active/active setup, you have to search for something 
else. Perhaps you need a cluster filesystem (ocfs, GFS, gluster or ceph). By 
the way, glusterfs has a NFS server included.


Mit freundlichen Grüßen,

Michael Schwartzkopff

-- 
[*] sys4 AG

http://sys4.de, +49 (89) 30 90 46 64, +49 (162) 165 0044
Franziskanerstraße 15, 81669 München

Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263
Vorstand: Patrick Ben Koetter, Marc Schiffbauer
Aufsichtsratsvorsitzender: Florian Kirstein
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 230 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20150420/7bc9f07d/attachment-0004.sig>