Can not mount NFS share with access restrinction

[Maxim Yakimenko]

Hi, collegues

I have two cluster nodes and a NFS-server.

I could manage to mount NFS-share to node where cluster packets are not active. But I can not mount same NFS-share to another node (with active cluster packets) - mount command says "Permission denied". In /etc/exports on NFS-server I specified access list using option -access=host1:host2.
When I erase access list from /etc/exports and reexport share I can mount it.

Any suggesions?

[Steven E. Protter]

Shalom,

Try using the IP address instead of the hostname in the access restriction. Also make sure the sytax of the access restriction is correct.

For an access restriction to work based on hostname, the hostname must resolve. I prefer to avoid that potential complication where practical.

SEP

[Maxim Yakimenko]

Well, I can ping second node from NFS-server by name and resolve it via nslookup, Besides, As I see access list in exports on HPUX nust contain only hostnames, it does not work with IP-addresses

[Maxim Yakimenko]

Syntax is identical on/for both nodes

[Dave Olker]

Does the NFS client that is unable to mount have multiple network interfaces? Is it possible the NFS requests are arriving on the server via an IP interface whose hostname is not in the access list?

Are both NFS client and server HP-UX systems? If so, you can collect a debug rpc.mountd log file during the failing mount and it might give you some hints about why the mount is failing. You could also collect a network trace of the failing mount request and that will also give you hints as to why the mount is failing.

If the NFS server is an HP-UX box you can send the running rpc.mountd daemon a SIGUSR2 signal (i.e. kill -17 ) and that will toggle logging to the /var/adm/mountd.log file. Reproduce the failed mount attempt, then send another SIGUSER2 to the rpc.mountd daemon to toggle logging off. Hopefully the logfile will tell you why the mount is failing.

Regards,

Dave