rpcbind daemon running at ninety-seven percent

[Andrew Kaplan]

Hi there --

We have an L2000 class server running HP-UX 11.00 and we detected the rpcbind daemon running at ninety-seven percent. This is causing users to be unable to 'lock' files via the system's exported NFS mounts.

What is causing this issue to occur, and what step(s) can I take to correct it? Thanks.

[TwoProc]

Andrew,

First, make sure you're patched up and have eliminated any bugs that may be in your server from this issue.

It's been years, but I've seen this wayyy back when we used to have HP workstations that would come in and out of the field, then hook up to the home office, get NFS mount points, and then depart without doing a proper unmount of the file systems, the systems would just be yanked from the network and carted off without proper disconnect procedures. Nothing I could do about it, except complain.

If this is possibly your situation, you can check this by doing a
$> showmount -a
command, if you've got the problem, you'll see lots of mount points to servers that aren't active, or maybe even not there at all. Another symptom that we saw was that the system was REALLY CHATTY, with lots of network I/O. Of course, back then we were 10baseT - so it didn't take much to consume the whole wire.

The solution was to go to single user mode, clean up some NFS related directories in /etc, and come back up. HOWEVER, that was HPUX 9.x, and I noticed that since 10.x - those directories don't exist or have moved, as I've not been able to find them. I even double checked my notes some time since, and wasn't able to locate the directories.

What it came down was the NFS services were desperately trying to maintain those connections to servers that no longer are on the network, some of them were gone forever.

Sadly, if this is your problem, I can't tell you how to fix it, but I'm sure a call to HP support would clear it up.

I also understand that the new NFS 4.x is much more tolerant of things like this, and would handle these types of events much better - but that is not available in HPUX 11.0.

Wish I could help you more.

[Dennis Handly · Accepted solution]

>TwoProc: The solution was to go to single user mode, clean up some NFS related directories in /etc, and come back up.

Are you talking about: (you don't need single user mode)
1) Kill rpc.statd and rpc.lockd ON BOTH SYSTEMS.
$ ps -ef | grep rpc.lockd
$ kill
$ ps -ef | grep rpc.statd
$ kill

2) Remove all entries in /etc/sm and /etc/sm.bak ON BOTH SYSTEMS.
$ rm -r /var/statmon/sm /var/statmon/sm.bak

3) Restart rpc.statd (first), then rpc.lockd, on both systems:
$ /usr/sbin/rpc.statd
$ /usr/sbin/rpc.lockd

Note: I haven't had to do these in years. Patches help.

[Andrew Kaplan]

Hi there --

I went ahead with the latter suggestion, and that seems to have solved the problem. Thanks for the help.

[Steven E. Protter]

Shalom,

If you fail to patch up the system, the problem is likely to return. Don't leave the job half done, finish it.

Further, HP-UX 11.00 is out of support. You might want to study getting that L2000 up to 11.11 for better long term support.

SEP

[TwoProc]

Yes Dennis,

That's it. Thanks for being able to complete my thought on how to do it. Way cool of you.

Thanks!