Rotation Stratergy

[bertieuk]

I am evaluating networker 7.1.

I have got the basic scheduling working but I seem to be having great problems on the strategy of cloning and tape rotation.

Environment: One Networker Server and one storage node. All clients backup to disk using advance file devices on either the Networker Server or the storage node.....This part is working fine.

I can probably keep upto 1 week of full and incremental backups on the disk storage (Approx 1TB in total). Each night I will group servers into fulls and incrementals and then rotate them. My idea was to have separate pools of full and incrementals to parallel stream to disk. I want to clone each night's backup onto tape and planned to use two tape drives/library. The full backup tape would be added to each day until the end of the month where it would then become the monthly archive and kept offsite for one year before being re-used.

My thoughts on cloning was to set the retention period for full backups to 1 year. I have set my full tapes to manual recycle and was going to use nsrmm to change the disk based backup to recycle after one week based on the backup date of the ssid.

By recycling the full backup do you automatically recycle the dependent incrementals?

Will the Full backups on tape be protected or will they be recycled?.....I wanted to keep the entries in the media index!!!

How can daily incremental tapes be reused without appending?

Is there a better way of doing this?.....I am using mminfo/nsrclone to clone to tape and the process has to be automated.

Would I be at an advantage if my Networker server was Linux and not windows?

Full clone tapes.....I planned to append cloned full, system files and index/bootstrap sets to the end of my full tape each day. Will the process of recovering from disaster be slowed down with all the different indexes being spread across the whole tape (most current being at the end).....could I do this better?

Thanks for the help Si

 

[Chapter11]

Saveset retention is based on ssid alone, not ssid/cloneid.

If you are backing up to disk, then cloning those savesets to tape, then setting the expiration of the saveset to 1 week, both the disk-based copy and the tape-based copy will expire after 1 week.

Using "nsrstage" will *move* a saveset instead of copy it, but you'll lose it off the disk volume.

I know this doesn't answer a lot of your questions. I found Networker to be weak (and confusing) in this area, and I've given up using its facilities for handling expiration. I export the media database to a set of mysql tables, and built a set of somewhere around 20 rules for defining the expiration of individual savesets, then I just feed that back into Networker via nsrmm -e I think.

 

[RMinNJ]

Without quite answering your questions I'd like to throw in
some thoughts:

Your using disk based devices. I assume your STAGING this
to tape at some point ? If you set the retention of a full
backup to 1 year, it will remain on your disk based device
for 1 year (if your not staging). As far as I know there
is no easy way to set the retention on the clones longer
then the retention of the original saveset.

If I have to do this I'd probably... simplify the retention
to be equal to the browse time that you want for all savesets .. say X months.
I'd set the retention to X months for everything. This will be enable the data on your disk device to come up for recycling..space will be made available for backups then.
Each night I would clone the data to an appropriate clone pool; Monthly Full Clones, Daily Incr Clones, etc. You
should be able to append any of these for 1 month..

Sorry if I misunderstood your setup.. I think its important
to note that even as clone tapes become recyclable (when the
original data becomes recyclable) the mediab still knows whats on the tapes and can recover from them (as long as they are not reused).