clone only full savesets

[spaceball]

Hi,
I've scheduled the backup as follows:
Filesystem backup:
savegroup: FILESYSTEM, schedule: Sun - Fr incr, Sat full
target pool is located on an adv_file

All FILESYSTEM Backups will automatically be staged towards a clonepool "FILESYSTEM_CLONE" on tape. (done by networkers staging-job)

Now I was thinking about using the auto-clone option inside the group-setup to just get a second clone only of the full backups:
FILESYSTEM should automatically clone towards a clonepool called WEEKLY, located on another adv_file

Clonepool "WEEKLY" should have a larger retention time and should only accept full level backups.

The problem is: "no selection criteria are allowed for clonepools"! It is not possible to reduce the clonepool "WEEKLY" only to full level backups.

any other idea?

 

[605]

Don't know why you do this. Although it is possible, it does not really make sense to clone from disk to disk, especially if this is all happening on the same storage node.
Instead, you should backup directly to this disk using the FULL pool (or equivalent) as a filter.

 

[a060463xyz]

To me its not really usefull but if you need the feature you have to implement a scheduled script or make an update to 7.6 SP1.

Here is the script :
Purpose cloning every full saveset taht is not older than 3 days and has not more copies than 2. (2 because of the adv_filedevices)

mminfo -r "ssid" -q "level=full,savetime >2 days ago, copies <3" >sidliste.txt
nsrclone -b "MyNewClonepoolName" -S -f sidliste.txt

Be aware that the cloning has to be done before the staging, else you might get two clonecopies.

 

[spaceball]

Hi,
to explain what and why I'm doing this:
to clone from tape (Clonepool "FILESYSTEM_CLONE") backwards to adv_file (Clonepool "WEEKLY") ist just a test.

As soon as I've enough tapes in my jukebox, this WEEKLY pool should take place on tape, as well as the FILESYSTEM_CLONE.

The reason for a second clonepool on tape and for cloning only the full backups additionally to clonepool WEEKLY is, that I have no alternative jukebox and I want to store a copy only of the last full backup (filesystems) on an offsite ocation. (let's call this "Disatser recovery for poor people")
I know that it might be easier to have a clone-copy at another storagenode or even another jukebox, but it is like it is.

So, again, my first idea was to automically clone -"postprocessing" via savegroup clone-config- and limit that destinating pool (WEEKLY)only to full backups (but that did not work) and put the weekly tapes offsite

and secondary stage (also auomatically by networker staging) towards the FILESYSTEM_CLONE pool, to have actual backups online available.

Is this really sooo nasty...?

 

[a060463xyz]

If you want to make an autoclone approach you have to clone everything, you can not limit the clonepools !

If you want to clone only specific savesets you have to use a sheduled script.

Thats the way it is !

 

[spaceball]

I think the situation has been cleared now: I made some scheduling- and savegroup changes and autoclone the full-backup towards an "offsite"-clonepool (tape). Than the normal auto-staging is moving the disk-backup towards the "normal" clone-pool (tape). So I have only one backup where only the full-levels went to two different clone-pools.

Thak's & Regards

 

[605]

BTW - i think NW's Automatic Cloning (7.6.1) would be ideal for your purpose.

 

[spaceball]

@605:
NW7.6.1 is not an option cause our customer uses FSC-Networker 7.5A00 and upgrading has not been ordered yet.

But in fact I realized it with networker auto-cloning:
-Assumed that the backup philosophy is backup 2 disk 2 tape.
-All clients have a scheduler "only_saturday_only_full"
(sun - fr: skip; sat:full)
and -for load-balacing all clients are fragmented into
several groups called
"fs_1", "fs_2"..."fs_n" for the full backups
and
"fs_1_incr", fs_2_incr"..."fs_n_incr" for daily incremental backups

-All "fs_n" savegroups are configured to use "auto-cloning " towards "weekly" tape-pool
-All "fs_n_incr" savegroups are using "forced incremental" and have dedicated scheduling "mo-fr_incr_sa-su_skip"
-All "fs_n_incr" savegroups are scheduled 10 minutes later than corresponding "fs_n" savegroup

So as a result:
-Monday to friday all savegroups "fs_n" will skip
-All savegroups "fs_n_incr" taken they're incremental backups onto disk
and will than be auto-staged towards "online-clonepool" (kept inside the library)
-Saturday and Sunday all "fs_n_incr" will be skipped
-and all "fs_n" savegroups taken they're Full backups
to disk and will auto-clone to "weekly-clone"
-than auto-stage will stage the full-backups to "online-clonepool"
-Monday: all online-available, used "weekly" tapes will be moved offsite

it looks a bit complicated but i guess that's it (!?)