I had problems with pstclntsave on one of my servers (NW6.1.1 on AIX
4.3.2). Best I can figure, it checked it's work list too soon and
decided the backup was completed and tried to run the post commands. I
finally got rid of the post commands and now run them as a separate cron
job well after the backup should be finished. It leaves my databases in
hot backup longer than they need to be, but it doesn't switch them out
of backup mode while the files are being backed up.
This happened quite a while ago, so it was probably running
6.0.something. I really ought to change it back and see if it's still a
problem.
Since this was on UNIX, I used my pre commands to start a background
process to "spy" on Networker every minute. I listed the networker
processes (ps -ef | grep nsr and ps -ef | grep save) and the work list (
echo "show work list\n print type: NSR group;\n name: mohawk_ora" |
nsradmin -s magic -i -). The worklist started empty (while the pre
commands ran), then showed some of my savesets, then was empty, then
showed more of my savesets, and finally was empty again. I'm still
spying on Networker but I don't see that pattern anymore. Like I said,
I need to switch it back and see if it's working properly now.
Your problem could be the number of savesets you are trying to do in
parallel. If the worklist for a particular client is temporarily
empties during the backup, the post commands will run. Then when
another saveset for that client is started your pre commands will run
again. It should run the post commands again when the worklist empties
again, but the timing of when the lock file is cleared gets involved
with this. That's the understanding that I came to when I was having
this problem.
Try adjusting paralellism and client priorities to make sure each
client stays busy saving without "taking a breather".
---
Jim McAlarney Huber Engineered
Materials
phone: 404-949-5505 4401 Northside
Parkway
fax: 404-949-1305 Suite 600
jmcalarney AT huber DOT com Atlanta, GA 30327
--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--- Begin Message ---
Hello all;
Server: NW v6.1.1 on a Compaq ML370 G2 within a Win2K environment.
We have 11 database servers (networker clients) which run in GroupA from
Monday to Thursday and GroupB on Fridays. The clients run cold every
Friday which requires that we use the savepnpc command (unless we
manually take down the dbs). For the past couple of months, the
pstclntsave.exe appears not to execute correctly, in fact not at all, on
10 / 11 of the clients. I was just curious to know if anyone else has
witnessed a similar experience.
This phenomenon just began out of nowhere with no warning. All Monday -
Thursday backups begin / complete successfully. I had first thought it
might be the <groupname.res> file. However, those files have all been
destroyed and created from scratch and tested successfully. The server
which succeeds appears to be random.
Appreciate any comments or suggestions.
Carter Fowler
Hardware Service Technician
The City Of London
600 380 Wellington Street
London, ON, CA, N6A 5B5
Off: 519.661.5711
Cel: 519.870.4956
Em: cfowler AT london DOT ca
--
Note: To sign off this list, send a "signoff networker" command via
email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--- End Message ---
|
