I have only a minimal knowledge of Linux/Unix so please bear with me.
I am trying to fix a problem at work. The problem is primarily with NFS services freezing up the system when it tries to load automatically. It may also be related to a possible problem with NISPlus and/or the system time.
This is a Redhat 9.0 system running the 2.4.20-8 kernal. As far as the services that may be involved, Redhat set the automatic startup priorities (run level 5) to 20 for autofs, 27 for nisplus and 60 for nfs services.
The first indication of a possible problem occurs when autofs starts. It indicates that it has started OK, but goes on to state: "Starting automount: name not served by this server. At this same point it used to state: "Starting automount: auto_home.org_dir: error in RPC subsystem". I'm not sure what caused the change. I still sometimes get the later error when trying to manually start autofs.
Next when NFS services try to start the system seems to stop. I then use CTRL-ALT-DEL to do a gracefull shutdown, boot in interactive mode and choose not to load NFS services.
When the system finishes loading Redhat frequently the time is off by about eight hours and sometimes it reverts to the previous day. I have the timezone set to Pacific (correct for this location) have reset the date/time frequently from inside Gnome and also, frequently I have reset the date/time in the BIOS setup program.I do not have the time service enabled.
This system is only an NISplus client. It is not an NISplus server. It is however a DHCP server, and holds most of the home directories for the network.
The last time that I manually got everything started I:
1. Stopped, then started the nisplus service
2. Stopped, then started the nfs service
3. Stopped, then started autofs
At that time these restarted with no errors. I know from past experience, if I reboot the system will "hang" again.
Any ideas?
Thanks in advance, for any/all help.
Well hmm, does your system behave when autofs and nis+ load at boot?
I wonder if something is taking too long to load, making nfs die. This sounds like what happened when I would LAN at my friend's house. I had in my fstab an nfs line that would stall my system. I ended up going into runlevel 1 and commenting it out. But that was after waiting over 30 minutes for it to stall (which it did not, I 3 finger saluted it to RL 6).
Hi,
1. Definitely fix the time issue. Implement ntp if you can. If this is a development server that people might build software on (or you might build software on) you'll save yourself a lot of hassle during the 'make' process. It's not clear if this is your biggest issue though.
2. When is 'portmap' started? Being that both of your mounting services are griping at you, this might be a good place to start.
3. Don't trust redhat's startup order. You probably need your services to start in the order a) portmap b) nisplus c) autofs d) nfs
4. Make sure you don't have some kooky iptables thing holding you back. Do 'service iptables stop' during your troubleshooting. Also check your /etc/hosts.allow and /etc/hosts.deny files.
Good luck.
[quote author=njcajun link=board=4;threadid=8210;start=0#msg76052 date=1072819468]
3. Don't trust redhat's startup order. You probably need your services to start in the order a) portmap b) nisplus c) autofs d) nfs
[/quote]
I doubt that is it, I ran them all just fine from RH 6.2 onward. I am sure one service is stuck and cannot continue. As to why ... well tthats what we need to figure out ;D
Posting Permissions
Reply With Quote
Bookmarks