[Lustre-discuss] Nodes claim error with files, then say everything is fine.

[Brian J. Murrell]

On Wed, 2008-08-06 at 11:08 -0600, Chris Worley wrote:
> 
> Would you suggest I increase or decrease this value?

Neither, just make sure all nodes have the same value.  I haven't seen
anything that indicates that it needs to be changed.

> Is there a way to inhibit the eviction,

No.

> or is that necessary to keep
> really dead clients from locking-out files.

Indeed.


> All the systems (RHEL4 and 5 clients, Lustre servers) are on the same
> ethernet and IB switches.  There were no issues before the 1.6.5.1
> upgrade with the RHEL5 nodes.

+rpctrace debug is probably the way to go to see what the client
are(n't) doing in terms of keeping the MDS aware of it's existence.

> Would a normal ping do it?

No.

> I can jury-rig all the RHEL5 nodes to ping the MDS.

Of course, even if it would work, this is a hack that would simply be
masquerading what is likely another, more real problem.  You need to get
to the root of the real problem.

b.

-------------- 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.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20080806/ce74c8d8/attachment.pgp>