From: Gary Lake on
We use staight NIS (no NIS+) for mail aliases and automount maps. We do
name resolution through DNS. Our NIS master is a Solaris 7 ( 5.7
Generic_106541-18 )
we've had a automount map that has been used to automount various
directories on different hosts ( Solaris 6 & 7 ) that have worked fine for
years. We also have multiple user home dirs so everyone get a mount of
/u/<username> as a home direcotry. Many of those home dirs on solaris 8
boxes. However we have another automount map called auto.stage2 that maps
several dirs on various sol_6&7 hosts. The problem is if I add a mount
point that points to any solaris_8 host it breaks all NFS exports. I've
tried everything from refreshing nfs/rpc/automountd daemons to a cold boot.
No matter what if the map points to a solaris_8 host things will work for
about two minutes, and then all exported NFS mount points break for the
solaris 8 hosts. The clients all respond with

The nfs client reports: <SERVERNAME> server not responding: RPC timeout in
it's syslog
The server that hosts the mount point reports no errors

Very strange. I thought perhaps things may be whacked because it was a 64
bit machine, so I tried pointing to a solaris_8 machine that I knew was 32
bit. No joy. I've played with timeouts, looked for errors in the automount
maps, tried EVERYTHING I can find in the NG's but no luck. I also looked
for speed/duplex mismatches. nisauthinfo shows only DES.

I can manually mount a directory exported on solaris 8 using either a
solaris 7&6 client, but if I change anything in auto.stage2 to point to any
solaris 8 host everthing goes wild.

Any Help Would be gladly welcomed.




From: bostontechgroup on
It sounds like you may need to upgrade to the newest version of autofs
on your older clients. Try that and let me know how it goes.

BTG

--
Boston Technology Group
http://www.bostontechgroup.com