make netbackup write to new media

[samar]

Hi all,

Netbackup 7.1 Enterprise.

I have volume pool with 24 volumes. Type of backup is oracle. rman starts 8 streams. So  it fullfill firs 8 volume,

and continue to write to second 8 media. and small part of backup is leaving on second part of 8 media. it has retention period 10 days.

and frequency is 1 week. So after week backup policy executes, and it continue to write on used media. That is also normal behavior of netbackup.

so it fullfill second 8 media and writes small ( but bigger then first time small part on first 8 volumes ) part of backup on third 8 volumes.

So after 3 days of second backup , retention period of first backup  is over and first 8 volumes wholy expired, and the small part of that expired backup on second 8 volume also expired. but because it writes linear I cant use that space until whole volume expired.

For first look it seems that is just small area, but after several cycle of backup it became like because of some little data on volume I am wasting

other big part of volume, and waiting for its expiration ..

Thats why I want netbackup to write on new media when it starts policy.

If it is possible I would give 16 volume ( with enough size) to this policy , not 24.

lets say it writes to first 8 volume. and that is ok if some small area leaves there empty.

and after week it starts to write to new media.after 3 days of this backup first 8 volume would be expired ..

Is it possible ??? or is there any way not to waste space on volumes. .. ???

Thanks

[Mark_Solutions · Accepted solution]

There are a few of ways around this ...

1. eject the tapes after each use so that they just get used the once (good practice to get them off site anyway)

2. Suspend the media after each use - that way they do not get writtent to again until the data has expired

3. Use vault to do both of the above (eject and suspend the media automatically for you)

Hope this helps

[Andy_Welburn]

Initially I would suggest separate schedules/policies to cover each period with an appropriate volume pool for each, but that may be a hassle to set up correctly initially (esp. with db agent backups)

Simplest way, but would require manual intervention by yourself or other admin, would be to eject or suspend those partially full media at the appropriate time until all images had effectively expired - but that would be a hassle on a regular basis!

***EDIT***

Mark's on the ball today!

[samar]

Thanks for reply, Andy, I also thought about as u said but there are too many policies, I wonder how to automate this. Mark thanks, i'll look to vault.

[Nicolai]

You may not like my answer but here it goes :D

This type micro management can steal all your'e time - and your'e time cost money also.

Just add tape free tapes so everything balance. For each hour you use for solving this issue you can buy 2-3 tapes.

Just a different angel.

Best Regards

Nicolai

[samar]

Nicolay thanks for reply, sorry missed one point to write. In my environment VTL, so in deep it is hard disk which is not cheap then my time:))))

[Marianne]

Thanks for telling us that your environment is VTL.

Seems your only choices are:

1) Manually suspend media after backups.

2) Separate volume pools with separate schedules.

[samar]

Hi all,

appreciate your replies.

so the solution was like as Mark said initially . Netbackup vault. great solution not only for situations

like mine but much more further.

Initially u need license for vault.

So I created vaults for each robot, and policies. skipped duplication and catalog backup of vault.

just like was using only eject feature of vault. but there I put deferred eject and enabled "Suspend media for the next session" and gave date interval like: suspend media on whic backups were written  between x and y days ago..  create vault type of policy, sheduled it after some time of vault  needed backup.

it started to vault, and actually what it did was just "eject" feature of netbackup vault, and it was not eject in real because of "deferred eject" option, it just suspended media, that was all I need ..

Thanks Mark a lot for correct reference,

But I got another , for my vtl, fujitsu cs800 i can create 9000 media per robot :)  , so here I got another solution, instead of creating 10 media each 1.5Tb , I created 150 media each 100gb .. which was really useful for me. I minimized wasting area on my virtual tapes.

Thanks all.

Good luck

[Mark_Solutions]

Glad it worked out for you - and on the VTL note you are right .. my last customer used 25GB tapes in their VTL which worked really well