And with subject, also...
______________________________ Forward Header __________________________________
Subject: No subject given
Author: LBRO <lbro_at_dscc.dk> at dk-internet-mail
Date: 7/17/98 1:37 PM
Hi, managers.
The problem, I reported was that in a cluster, I suddenly got a lot of
errors in the uerf log referring to disks that other members of the
cluster had mounted.
Almost all agreed that this was caused by 'advfsd' that monitors care
of the file sets and domains. This program makes a sort of 'scu show
edt' and monitors all the disks found here. But some of those are
reserved by other members of the cluster and that causes the error.
Chris H Ruhnke suggested that one could put the names of the disks
mounted by other cluster members into a 'ignore' file (THANKS!!, it is
not documented in the man pages) and then restart advfsd. That seems
to work quite well with the following to note:
On my system (DU4.0D patch kit 1):
It is /var/opt/... instead of /usr/opt/...
It doesn't seem necessary to restart advfsd.
I attach an awk script that I have made that can be run by cron. It
makes an ignore file that contains those disks of the disks visible by
scu that are not below /etc/fdmns.
Thanks to
+ Chris H Ruhnke who suggested changing /usr/opt/advfsd/disks.ignore
+ Mike Cross
+ Viktor Holmberg
+ Steve Parker
+ rye_at_jtasc.acom.mil
+ Rodrigo Poblete
+ Ken Krueger
Received on Fri Jul 17 1998 - 13:48:20 NZST