3003 networking problems
Ruth Milner
rmilner at zia.aoc.nrao.edu
Wed Apr 17 08:39:15 AEST 1991
Our systems were just upgraded to 3003 (by the local IBM office; they wanted
our demo loaner and our own to be at the same level for debugging purposes)
and I have noticed two rather nasty problems with NFS and NIS (aka YP):
1. The "bg" option on NFS disks in /etc/filesystems no longer works. If a
remote server is alive at boot time, the disk mounts just fine. If it
isn't, the timeout and backgrounding appears to happen normally (usual
delay and messages), but in fact it just vanishes. The disk never gets
mounted after the server comes back, and when you look at system processes
there is no "mount" hanging around. The disk has to be mounted by hand.
When you mount half-a-dozen data/code areas from other systems, this is
annoying, especially if it happened because of a power failure in the
middle of the night.
2. Both of our RS/6000's are NIS clients served by a Solbourne running the
equivalent of SunOS 4.0.3. Both were fine at 3002, all our other systems
are fine, but now at 3003, occasionally ypbind on AIX goes catatonic. This
is a problem because we only use the IBMs as compute engines; most users'
home areas, all our group and passwd information etc., is obtained through
NIS. So when this happens, all logins not explicitly in /etc/passwd hang.
It is easily fixed by killing ypbind, then srcmstr automatically restarts
it, and the hung logins magically resume. But it is a real pain to have to
do this.
Has anyone else seen either of these problems? Both have happened several
times since the upgrade three weeks ago.
I called both of them in to the Austin hotline. The fellow who called me back
on the NIS problem sounded reasonably competent, but despite having left a
brief description of the NFS problem, the woman who called me on that one not
only didn't know what the "bg" option did, she didn't even know what the
/etc/filesystems file was for! Hey, IBM, it really isn't a good idea to have
people respond to calls when their area of specialization lies somewhere else
entirely. It's not exactly reassuring to have to explain the problem at that
low a level.
Thank you.
--
Ruth Milner
Systems Manager NRAO/VLA Socorro NM
rmilner at zia.aoc.nrao.edu
More information about the Comp.unix.aix
mailing list