1) Putting the backup server into each group with the clients appears to
work (based on one test).
2) We thought about putting all the index information into a separate
pool, but we were concerned about one tape holding so many indices. To
minimize this risk, we were thinking of writing a script to rotate
through a set of tapes. For example, if 2 nights of index backups fit
on one tape, write:
tape1, tape2, tape3, tape4, tape1, tape2, tape3, tape4
tape1, tape1, tape2, tape2, tape3, tape3, tape4, tape4
To do this, we'd mark the tape that was written yesterday as readonly
and then the next tape in the chain as appendable.
3) We thought of this too, but are a little hesitant to alter the
retention times with a script.
Is having 3 definitions for each client clumsy? Well, not after they
are set up - and we have a script to define new clients. It walks the
user through entering name, group, etc. and then creates appropriate
nsradmin input files. Also, so far, maintenance is not too bad. It
does cause a couple issues, however, including:
- When we use owner notifications, multiple messages are created. We
have a script which consolidates messages for a group of clients before
sending out the notifications (most of our owners are interested in
- When the bootstrap is printed, we get multiple copies.
We like this configuration because it's using Networker to schedule
Networker tasks. That is, the control is all within the software rather
than distributed out to a variety of scripts. When designing this, it
seemed simpler to keep all the control in one place. This set up is
fairly new in our environment, but so far, it's working well for us.
From: Davina Treiber [mailto:Treiber AT hotpop DOT com]
Sent: Friday, February 27, 2004 3:24 AM
To: Legato NetWorker discussion; Teresa Biehler
Subject: Re: [Networker] setting different browse policies
Teresa Biehler wrote:
> I have a need to have different browse and retention policies for
> weekly and monthly backups. For each client, I've defined:
> - a definition for dailies with a browse/retention policy of 5
> - a definition for weeklies with a browse/retention policy of 3
> - a definition for monthlies with a browse/retention policy of
> one year
> The different types of backups are filtered into different pools by
> The data save sets are recycling as expected, but the index save sets
> are picking up the browse/retention policy for the server which also
> the above browse/retention policy. It appears that all index savesets
> are using the browse/retention policy of one year. This is causing a
> problem because the daily pool is supposed to recycle tapes after 5
> weeks. Since the index save sets are written to the same tapes and
> have a browse/retention of a year, this is not happening. How can I
A few options:
(1) I _think_ that if you put the NetWorker server in the same groups
(perhaps just backup a single file) it will sort out the retention for
the index save sets.
(2) Backup index save sets to a separate pool. Plenty info in the
archives about how to do this.
(3) Run a script (weekly?) to change the expiry date of all save sets in
a pool to a specified offset from the save time. I can supply you such a
script off-list if you like.
Do you have three client resources for every client? This can get a bit
clumsy can't it? I have one client resource for each but in three
groups, using the group schedule to override the client schedule.
Unfortunately there is no group override in the GUI for browse and
retention, even though it is available as an option to savegrp. To fix
this I back them all up with browse and retention of "Year", then change
it using the aforementioned script. Works well.
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.