start NFS core sub system

[Eric Zumwalt_2]

Hi,

Having a problem when I rebooted my HP-UX 11.0 server. When it goes through the start up processes it freezes at the, start NFS core sub system.

Any ideas why this is happening?

Thanks,

Eric

[Bill Hassell · Accepted solution]

This is very common problem. The only way to fix it is to turn off NFS and reboot. You may not be able to break out of the NFS lockup so you can see if you can telnet to the server. Edit the file: /etc/rc.config.d/nfsconf and set all the NFS services =0.

Otherwise reboot into single user mode. Since single user mode does not have /usr, /tmp and /var mounted, you'll need to mount those 3 directories. Then edit the nfsconf file and reboot.

[Dave Olker]

Hi Eric and Bill,

Bill - I had no idea this was a common problem. I've never heard of it before. I've seen stops responding in nfs.client (when there were problems with automount/AutoFS) but I hadn't heard of a common problem with stops responding in nfs.core.

I don't see how setting NFS_CLIENT or NFS_SERVER in the /etc/rc.config.d/nfsconf file will change this behavior, since the /sbin/init.d/nfs.core script doesn't source the nfsconf file, nor does it make any decisions based on the value of these NFS_* variables.

All nfs.core does is take care of some backward compatibility stuff (i.e. adding entries in /etc/netconfig and /etc/net/* directories, create device drivers for tlcots, tlcotsod, and tlclts) and then start the rpcbind and nfskd daemons.

Again, I haven't heard of any common problems with any of these steps. However, I've been out of WTEC for a few years, so perhaps this is a common problem.

In any case, I'd be curious why it is happening.

One way to find out where the nfs.core script is stopping would be to boot the system in single-user mode, mount all the local filesystems in /etc/fstab, and then modify the nfs.core script by adding a "set -x" command to the start of the script so that each command in the script is printed to stdout. Then run the script and see where the script stops  if you're willing to do this, please let me know where the nfs.core script is stops responding .

Regards,

Dave

[Dave Olker]

Hi Eric and Bill,

Bill - I had no idea this was a common problem. I've never heard of it before. I've seen stops responding in nfs.client (when there were problems with automount/AutoFS) but I hadn't heard of a common problem with stops responding iin nfs.core.

I don't see how setting NFS_CLIENT or NFS_SERVER in the /etc/rc.config.d/nfsconf file will change this behavior, since the /sbin/init.d/nfs.core script doesn't source the nfsconf file, nor does it make any decisions based on the value of these NFS_* variables.

All nfs.core does is take care of some backward compatibility stuff (i.e. adding entries in /etc/netconfig and /etc/net/* directories, create device drivers for tlcots, tlcotsod, and tlclts) and then start the rpcbind and nfskd daemons.

Again, I haven't heard of any common problems with any of these steps. However, I've been out of WTEC for a few years, so perhaps this is a common problem.

In any case, I'd be curious why it is happening.

One way to find out where the nfs.core script is stopping would be to boot the system in single-user mode, mount all the local filesystems in /etc/fstab, and then modify the nfs.core script by adding a "set -x" command to the start of the nfs.core script so that each command in the script is printed to stdout. Then run the script and see where it stops responding i.

Eric, if you're willing to do this, please let me know where the nfs.core script is stops responding i.

Regards,

Dave

[Dave Olker]

Sorry about the duplicate posts, I've been experiencing major connectivity problems with ITRC tonight.

Dave

[Bill Hassell]

This is a 'feature' in 11.0 and 11i core installs only. It is probably due to partially completed network configs, misconfigured DNS, etc, but check with Randy S. in Roseville or some of the sysadmin RCEs for some history (or KDB). I can duplicate this on an 11i install by changing just the hostname in netconf but not the IP and using nsswitch=files the dns.

[Jan Sladky]

Hi Eric,

Bill is right. Additional info - you should be able to interrupt NFS core "freezing" by folloving :

ctrl+\ - jump over
trl Break - jump inot console

than repair your NFS related settings

br Jan

[Eric Zumwalt_2]

Thanks for all the reply's. I'll try this out and let you know.

Thanks,

Eric