nfs "unreadable" error

[Erol KAHRAMAN]

hi,
I have two hp-ux box. Each of them share one directory to other server with nfs(automount).
SRV1
/etc/exports
/dir1 -anon=65534

/etc/auto_master
/- /etc/auto.dir2

/etc/auto.dir2
/dir2 -rw SRV2:/dir2

SRV2
/etc/exports
/dir2 -anon=65534

/etc/auto_master
/- /etc/auto.dir1

/etc/auto.dir1
/dir1 -rw SRV2:/dir1

My problem is;
When i restart my servers or one of them my shares appear unreadable.
When i execute "exportfs -av" on servers all is going ok.
What could be a problem ?

[Berd]

Erol,

I would suspect you have a stale NFS mount. As when the server(s) are rebooted the exported directories are unavailable. Do you see stale entries when you bdf ?

Regards,
Berd

[Dave Olker · Accepted solution]

Hi Erol,

What version of HP-UX are you running and which version of AutoFS are you running?

Are you able to manually mount these filesystems between hosts? Take AutoFS out of the picture and first make sure manual mounts work as expected. Once you verify manual mounts we can put AutoFS back and see how it goes.

Also, looking at your maps:

SRV1
/etc/exports
/dir1 -anon=65534

/etc/auto_master
/- /etc/auto.dir2

/etc/auto.dir2
/dir2 -rw SRV2:/dir2

SRV2
/etc/exports
/dir2 -anon=65534

/etc/auto_master
/- /etc/auto.dir1

/etc/auto.dir1
/dir1 -rw SRV2:/dir1

It looks like both of these mount points are served by SRV2. Is this true? Is SRV2 doing a loopback mount to itself? Which of the mounts works and which one fails? Is SVR1 able to mount from SVR2 but the mount on SVR2 fails, or vice versa?

Regards,

Dave