Tape jobs

[Alvaro]

Hello

I read in the manual this:

''When you add a repository as source to tape job, the tape job constantly scans the selected repository (or repositories) and writes the newly created backups to tape. The tape job monitors the selected repository in a background mode. You can set explicit backup windows for the tape job. In this case, the tape job will start on the set time and archive all new restore points that were created in period since the last job run.''

My question is...If I create a restore point full backup in my repository...when the tape job scan the backup chain in the repository ..does it copy the full..or does it copy an incremental forever?? I am asking this question because I know that copy jobs always make incremental forever..doesnt matter if I have a full in my primary backup ..and I would like to know if tape job could do this.

this is important because I need to know the amount of data to size the number of drives in my library

Thank you

[HannesK]

Hello,
Yes, it will copy the full backup file.

I'm sure you don't want to run tape jobs with "incremental forever". Imagine you have to restore hundreds of incremental backups first...

Best regards,
Hannes

[Alvaro]

Ok

I need to copy my primary backup which is on disk to secondaty destination in tape.

For this I have to run tapes job. The tape job can use two different sources to copy:
-The source repository
-Backup job or copy job.

In the first case of the source repository, As you told me the tape job copies exactly what appears in the source, that is, if I do a full job, the tape job will copy a full one. But I have the problem that this client has a lot of data, and a full one would take a long time to go to tape and I would need a lot of drives.

So I thought about using copy jobs that always act incrementally as a source. In other words, if my client makes a full in its primary chain, the copy job will do incrementals (not matter if my customer makes full) and stores it on an intermediate disk to wait for tape job. The incremental created by the backup copy job is then copied to tape. Is this approach correct? I need to reduce the backup traffic to tape ..since my client use forward incremental (not forever) and the day when He make full..I need to put a lot of drives..

Thank you a lot

[HannesK]

there seems to be a misunderstanding: there is no way to make a tape job to work like a backup copy job. there is no way to make tape jobs "incremental forever". that would result in horrible restore scenarios.

At some point you need to write a full backup to tape again (usually weekly).

Incremental forever works fine on disk because we can do merges and there is random IO on disks. Tapes work sequential. That means no merges and no random IO.

The only way to reduce a little bit of traffic is by not backing up incremental files at all.

[Alvaro]

Of course I know that doing incremental forever without any full is a crazy thing but I am not speaking about that.

I know that copy job always produces incremental restore points, never fulls.

So....
According to the manual I can choose a job (Backup job or copy backup job) or a repository as a source of my tape job.I read in the manual the following:

If the source job is backup copy job, keep in mind that the last restore point of the backup copy job stays active until the next restore point is created. The tape job does not copy such active points, because they may be updated. For this reason, the backup chain on tape will be always one restore point shorter than on disk.

When we have repository as a source:
The tape job monitors the selected repository in a background mode. You can set explicit backup windows for the tape job. In this case, the tape job will start on the set time and archive all new restore points that were created in period since the last job run

So, Now what I have is a customer who is making forward incremental (One full every week) in his primary backup chain. After that I could lauch copy jobs which make a forever incremetal copy on disk. So If I choose the copy job as a source..I will have incremental restore points which I could copy to the tape with tape job. Of course..After I would do virtual full on tape regularly to have better restore scenario...because having only incrementals is crazy.

On the other way..If a choose repository as a source the tape job copy exactly the primary backup chain...and when the customer make a full of 357 TB is impossible to send this amount of data to the tape in good interval of time..I would need a lot drivers..

For that reason What I am asking is If the first option (copy jobs create incremental on disk and after I use tape job to send them to tape) is possible and a correct way to decrease the traffic of amount of data.

Thank you so much

[HannesK]

first option (copy jobs create incremental on disk and after I use tape job to send them to tape) is possible

yes you can do that. you already mentioned virtual synthetic fulls: that's the way to go.

If you create the virtual synthetic full less often than once a week, then you will save space on tape. And if you have several BCJs as source, you could do the virtual synthetic full backups on different days to "load balance" the amount of data.

[Alvaro]

Thank you Hannesk for your quick responses! Very usefull!!

Best regards!