hanfs problem(server retries for nfs server package)

kirit_2 · ‎11-18-2004

first my setup is like ...

two nos of rx2600 server with 11.23 OS and sept'2004 patch bundle is loaded on both.

SGeSAP and HANFS is configired for high availbility using two package concept.

one package name ci runs on server A
Second packgae name db runs on server B
db packgae has nfs exported file systems ( /export/usr/sap/trnas and /export/sapmnt)
and these are exported throug hanfs on server A as well as server B as /usr/sap/trans and /sapmnt.

My problem is that when I shutdown the cluster , some commands hangs and console gives messages that "nfs server db not responding still trying"

can u please guide me what could be the problem?

Uday_S_Ankolekar · ‎11-18-2004

This doc may help:

In a NFS Serviceguard environment the client side mounts, autofs or static,
will go stale when the package is moved on the NFS Serviceguard server from the
Primary node to the Secondary node.

Why?
RESOLUTION

Stale NFS mounts on any NFS client are caused when the NFS inode is changed
while the mount is active.

When preparing to confiure a NFS package the names of the volume groups must be
unique within the cluster, and the major and minor numbers associated with the
volume groups must be the same on all nodes. In addition, the mounting points
and exported file system names must be the same on all nodes.

The preceding requirements exist because NFS uses the major number, minor
number, inode number, and exported directory as part of a file handle to
uniquely identify each NFS file. If differences exist between the primary and
adoptive nodes, the client's file handle would no longer point to the correct
file location after movement of the package to a different node.

This is what results in the Stale file on the NFS client when the package is
moved.

To check the major and minor numbers for your NFS package filesystems you can
run the following command.

# ls -al /dev/*/group

For more details on Configuring the NFSTOOLKIT package please reference the
following manual:

Managing MC Serviceguard NFS

http://www.docs.hp.com/hpux/onlinedocs/B5140-90017/B5140-90017.html

Good Luck..

kirit_2 · ‎11-18-2004

Hi uday

here nfs exported file systems are part of db package only.

Checked minor number configuration also , its ok.

also I get error even when I stop the cluster ( both packags show down in cmviewcl)

Dietmar Konermann · ‎11-18-2004

kirit,

You are saying that both packages are in status "halted"... so the package control scripts themselves are not hanging, correct?

I think, what you are seeing is quite normal. SGeSAP does not umount the NFS mounts during package halt. This means that, as long as the NFS serving package is halted, the correpsonding (relocatable) NFS server is unreachable. And since these mounts are usually hard mounts, some commands may hang and console message may appear.

What is your actual problem? If you dislike that behaviour during e.g. maintenance then maybe a dedicated NFS package (which is kept running) could be an option.

Best regards...
Dietmar.