Hello Managers:
I have been searching the archives for related dialogue and have had no
luck on gaining clues to my problem, so perhaps you may have some helpful
suggestions for me, or point me in the right direction.
Background:
I am very new to the Digit.. err.. Tru64 Unix advfs. I inherited the
system from a systems administrator that quit to return to a life of
farming so I received very little tutoring.
The system is running Unix v4.0b. To my knowledge there have been no
patches applied.
I have been running level 0 vdump weekly and level 9 nightly on 6 advfs
file set clones.
These vdumps have been running smoothly for 2 months.
This is not a 24x7 mission critical system by any means. It is for
development and testing and is therefor running a few applications.
Problem:
In the last 2 weeks, the backups have been hanging my system, requiring a
reboot. The scripts for the level 9 and level 0 vdumps are run from the
crontab file at 21:00. The scripts create all the clones up front for the
6 file sets. Then individually, each file set clone is mounted, vdump(ed),
umount(ed) and removed before continuing to the next file set clone. So
far the system "hang" condition occurs during the vdump of the third clone.
The first clone dumped is / (root), the second is /usr and the third,
/home, is a relatively large file set with user data and application
programs. The vdump of this third clone begins approximately 3 minutes for
level 9 vdump after the clones were created.
No errors are logged, disk space is not full on /home (a bit tight, but not
full) during the event. The only other information I think is pertinent is
one application that executes and writes log files to this file set is
running through the evenings now all week long.
I can run my scripts interactively with no problems. I can execute them
from cron during the day with no problem (while the application is not
running). Only after 21:00, while I am at home relaxing, right?, does the
system hang. By "hang" I mean all interactive seesions are lost, programs
stop running, console is unresponsive, system is not pingable on the
network, etc. so the reset button must be used.
Question:
Is there a known problem with using clones during periods of heavy usage
with this version of O/S? Or is this even related. Do I need to start
applying some patches?
Thank you in advance for any insight or direction!
Sincerely,
Steve Shanks
-----------------------------------
Steve Shanks (shanks_at_probita.com)
Probita, Inc.
4949 Pearl East Cir. Suite 300
Boulder, CO 80301
(303) 449-7665
Received on Wed Apr 21 1999 - 16:44:10 NZST