ADSM-L

Re: TSM Central Scheduler Stops Prompting

2004-08-23 18:58:41
Subject: Re: TSM Central Scheduler Stops Prompting
From: David E Ehresman <deehre01 AT LOUISVILLE DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Mon, 23 Aug 2004 18:58:55 -0400
When it happened to me on AIX 5.1, TSM 5.1 it was a TCPIP Receive Buffer
error.

(echo "----";date; netstat -v ent1 | grep "Receive Pool Buffer" )

where ent1 is your ethernet interface will tell you if you have receive
buffer errors. "No Receive Pool Buffer Errors: " should be zero.  If
not, you need to reset the size of your Receive Buffers.  Mine are set
at 2048.

David Ehresman
University of Louisville

>>> Curtis.Stewart AT LAWSON DOT COM 08/23/04 4:43 PM >>>
Has anyone run into similar issues with TSM 5.2.2.3?

HOST OS: AIX 5.2
TSM Level: 5.2.2.3
Server: RS/6000 P-630 dual CPU 12GB RAM
Two TSM instances running. The system is not LPAR.
One library manager, one client. Library sharing through TSM.
Library: 3494 with 8 3590 drives, direct connect SCSI.

It's the second time this has happened to me with the current system. I
have two instances, one has about 3 times the clients of the other, but
backs up roughly the same amount / number of files a night. After some
time, roughly a month or two, the prompter has just stopped prompting.
The
only clients that ran are either in polling mode, or run via some other
scheduler. It happened again to me this weekend, and I just happened to
log in and catch it. I restarted TSM and ran immediate action schedules
for the 87 missed clients. The rest of the backups that were scheduled
to
start later in the evening finished fine.

I looked in the actlog and the only errors where the standard "Client X
missed it scheduled backup" type things. There weren't even any messages
about the system attempting to contact a client for backups. I looked at
client side dsmsched.log and dsmerror.log files and there were no
entries.
There's nothing in the AIX errpt either.

I recall having this issue once at I think version 5.1.6.5, but can't
recall what the fix was. I think it had something to do with TCP/IP. But
don't have the documentation from the previous job.

I did find this thread on the adsm.org website. But nothing else.

http://msgs.adsm.org/cgi-bin/get/adsm0303/959.html

curtis.stewart AT lawson DOT com

<Prev in Thread] Current Thread [Next in Thread>