Networker

[Networker] SUMMARY: Retention policies revisited (need advice)...

2003-07-21 15:26:54
Subject: [Networker] SUMMARY: Retention policies revisited (need advice)...
From: Craig Ruefenacht <Craig.Ruefenacht AT US.USANA DOT COM>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Mon, 21 Jul 2003 13:24:53 -0600
Hi,

I want to thank those who sent replies to my post about our retention policy
problem.  Using that information, I spent some time Friday and over the
weekend testing some things out and was able to get retention policies to
behave themselves.

On Friday I setup a test Networker group with a test client and did some
backups to a test tape pool.  My first test showed that the retention policy
wasn't following what it should be, so I deleted the test group, test
client, and test tape pool altogether and then started over.  After that,
the retention policy started working as it should.

To fix my real Networker groups, I removed the groups (but not the clients)
and then recreated the groups and put my clients back into the new groups,
and over the weekend, things worked as they should.  I am not sure why this
would have fixed my problem, but it appeared to do so.

In summary, retention policies for client indexes behave like this:  In an
active Networker group, if the Networker server is listed as one of the
clients in the group, the client indexes that get written to tape will have
the same retention policy as is defined for the Networker server's client
entry in the same group.  The Networker server client entry in the group can
have a schedule that skips on all days, so when the group runs, none of the
filesystems on the Networker server get backed up - except for the client
indexes.  Having a schedule that skips all days for the Networker server may
be desired in some Networker groups where it isn't desired to actually
backup the Networker server using the group, i.e., the only reason for
having the Networker server listed as a client of the group is to set the
client index retention policy for the group.

In an active group that does not have the Networker server listed as a
client, the retention policy appears to be set by taking the longest
retention policy specified in all the Networker server client resources.

For instance, if you had 10 Networker groups and the Networker server was
listed as a client in 7 of those groups, the retention policy for the client
indexes in each of those 7 groups will be the same as the retention policy
belonging to the Networker server client in each of those groups.  In the 3
groups that do not have the Networker server listed as a client, the
retention policy is found by looking at the Networker server client
resources and taking the longest retention policy found.

I also read through parts of the Administrator's Guide for Networker 6.1,
Unix Version, page 206, where it talks about directing the client indexes
and bootstraps to a different pool.  At first I thought that this was my
answer.  I attempted to follow the instructions with no luck.  It appears
that the instructions contained in the manual are lacking in some details -
I am sure one could get it to work as advertised, but there must be some
other things that need to be tweaked in order to do so.


--
Craig Ruefenacht
UNIX Systems Administrator
USANA Health Sciences, Inc.
http://www.usanahealthsciences.com



--
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.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

<Prev in Thread] Current Thread [Next in Thread>
  • [Networker] SUMMARY: Retention policies revisited (need advice)..., Craig Ruefenacht <=