|
Re: [Networker] networker won't shutdown
2005-10-29 05:16:31
Joel Fisher wrote:
It actually is a pain to manage, but I've just become fed up with one
client that hangs an entire group and then it doesn't run the next
night.
Since I have an entry for the server in each group(for index retention
length purposes) it does actually put quite a load on the server. I'm
trying to decide what I'm going to do about that. It runs 600+
bootstraps a night which cause the server to not respond at times. I'm
considering moving the indexes to another pool so that I don't have to
worry about the retention time causing my media to not recycle.
By your response, I take it you use large groups. Do you not have
problems with hanging clients? How do you handle your indexes?
This must be a management nightmare. This solution is like a
sledgehammer to crack a nut, and as you have discovered, has introduced
more problems as side-effects.
I would have chosen to solve the problem in a much more simple way, by
using a simple script to kill off any overrunning groups. I believe I
posted such a script recently. Even something that alerts you to
long-running groups each day would be a good solution. I have used both
of these approaches for different customers.
Given that you have taken this step though, you have a real problem with
bootstrap backups. Running 600+ bootstrap backups a day is a very bad
thing, mainly because when the bootstrap backup runs it shuts down the
media database for a couple of minutes (YMMV). (This is why you see the
server becoming unresponsive. I don't know whether or not you run RMAN
backups but in my recent experience, a boostrap backup is likely to kill
a running RMAN backup. It is not easily possible to prevent bootstraps
from being saved, the only way I know is a bit of a kludge where you
move out the save command on the server and replace it with a script
that selectively does not save bootstraps, but allows other backups
through. Obviously it needs to save the bootstrap sometimes, but the
script can allow for this in a specified case.
You could implement such a solution, but this would be like (another
cliche) the old woman who swallowed a fly, and I would advise you to go
back to multiple clients in a group.
It is often sensible to take the index backups out of the normal groups,
and run them all in a single group (savegrp -O) to an index pool. This
certainly seems like something you should do here, whatever group
strategy you are using.
I really believe you have taken the wrong approach here, there are much
easier solutions to your problem. I will email you a couple of scripts.
To sign off this list, send email to listserv AT listserv.temple DOT edu and type
"signoff networker" in the
body of the email. Please write to networker-request AT listserv.temple DOT edu
if you have any problems
wit this list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|
|
|
