Clone immediatley, but not removing from disk storage until space is needed

It all comes down to how much busy your disk and drives are and that would determine approach, but in short you will need 2 scripts. Let's assume you have backup window over the night, then you would have cloning window in morning and staging window in afternoon. Call scripts from cron and that should keep your environment happy.

More or less correct


...where watermak can be either time data spent on disk or how of disk in total in being used (whatever comes first).


You lost me here as then I wish to ask you what is the purpose of cloning for you in the first place? I see no problems with staging after....


That would backup to disk.

That would be restore from disk.

That would be cloning via either script or automatic one from group (even that one is not very smart, but for cloning from disk it does not matter in reality).


Actually that can be done automatically.


I think you missed few important things here. As it seems to me, you do not wish to have data staged as data is already cloned to tape, right? fyi, in NW world instance on disk and instance on tape are not the same so NW will never do what you want NW to do it. Further, 2nd copy (cloning) should provide security. If you don't wish that then simply skip cloning and use staging only (or have b/r small as your retention on disk is and have clone with higher retention even I see that as complication that should be avoided).

Thats a good start, but its not really what I'm looking for.

As it appears now, networker's built in cloning and networker's built in staging work without any shared logic.

You can choose:

1) To clone immediatley after a backup.
2) To stage when the watermarks are hit, freeing up disk space.

What I'd like to do is:

Have it clone immediatley after a backup, but leave the data on the disk array until the watermarks are hit. Normally at this point, it would stage the data, but since it's already been cloned off to tape, this is a redundant procedure.

The goals are:

1) To speed up backups since the tapes don't get involved until the backup is done
2) To speed up restores of recent data, because its not removed until the space i needed
3) To get the data off to tape as soon as possible(immediate cloning).

Currently it appears the only way to get the data off a disk array is to stage it off, manually remove it or expire it and let networker reclaim the space. I'm wondering if theres any way to make the staging process recognize that its already been cloned off and just reclaim the space when needed.

I guess I'm doing a pretty bad job communicating what I'm hoping to accomplish.

Basically I'd like to:

Clone immediatley after backup (I know how to do this)
Leave the data on the disk array until the space is needed, and then have it automatically purged. The only way I know how to do this is via staging, which in effect, clones it a second time. What I'm trying to avoid.

OK. For the sake of audience let's add that cloning saveset is creating a copy of original saveset.


Wrong. Staging does not create second clone. Staging moves data from source to target. So, cloning is copying and staging is moving. Since you wanted to create a copy in first step with clone I can't see the reason why do you wish to delete original one now?

You can do that, but it takes script of course. Write script that will same check as staging does and then remove those savesets from disk and call for cleanup.

Semantics arguments about what staging/cloning do aside:

Intially, I'd like to make an immediate clone, because it clones to an offsite storage node, providing a backup outside of the server room in case of fire/natural disaster. At this point I have a readily available copy on the disk array for normal restores and the offsite clones in the event of a large scale disaster.

I'd like the data to persist on the disk array till the space is needed for further backup, and then be purged from the array, much the way the watermarks for staging policies do. At this point, the backup is only needed for large scale disasters, so the original offsite clone is sufficent.

Staging policies already have the logic to recognize that the array has passed a certain use threshold and remove data from the array in an intelligent manner(based on age, size, etc). Is there a way to access this logic from within the normal networker console without requiring it to first replicate off to a another device? This is the information I'm looking for.

If this is impossible, I'm looking for suggestions on how I might access or duplicate this logic in a script, taking into account the current usage of the source array and the existence(or non-existence) of clones of the to be purged data.

Hi Vic, that won't go as both steps are seen as part of single stage session. The only way to lock ssid on device is to trigger read operation that will stop/block delete operation, but it will be executed upon next nsrim or recover space operation.

Vic,

This is exactly what I'm trying to do.

Thanks,
-Tony

No.


Do a script which:
a) checks for disk usage and free space availability and checks against your threshold
b) lists ssid/cloneid pairs on device based on savetime and their size
c) calculates ssids and their size until low watermark you set in reached
d) calls nsrmm and deletes selected ssids/cloneids
e) calls nsrim check for device to make cleanup