Our SP/2 configuration has 5 nodes (three high and two thin nodes) connect
with a High Performance Switch.
We use ADSM to backup to two 3590 tape devices. When we run the backup
scripts in parallel the error occurs.
The ADSM-Server is a high node SMP 604-4 with 1 GB memory, and 4 processors.
Backup is taken via the HPS.
Karl B. Molter wrote:
Dear ADSM Experts,
I have running ADSM V 2.1.05 on an IBM 43P Server (AIX 4.1.5)
Client Version 2.1.08
I occasionally realize failing incremental backup schedules on clients
in
subnets different from the server subnet.
The log file on the client shows:
03/10/1998 21:43:02 Retrying scheduled event: DAILY-INCR
03/10/1998 21:44:17 ANS4017E Session rejected: TCP/IP connection failure
03/10/1998 21:44:17 ANS4847E Scheduled event 'DAILY-INCR' failed.
Return code
= 1.
I do not find any corresponding error-log entries on the server or client.
There are no interface errors (fast ethernet on the server, built in
ethernet
on the client) too!
This problem never occurs with clients in the server subnet and only
occasionally happens with clients outside!
I have any night running about 50 client backups with 0 to 2 failures
of the
above kind.
I do not have any explanation for this behavior.
Has anyone of you realized something similar?
Best regards,
Karl Molter
--
Dr. Karl Molter
Voice: +49-651-8103-215
Fachhochschule Trier
Fax: +49-651-8103-214
Computing Centre
Mail: K.Molter AT fh-trier DOT de
Schneidershof
WWW: http://www.fh-trier.de/~molter
D - 54293 Trier / Germany