Backup job with 1 restore point not merging..

[Dave338]

Hello, I'm in the process of making a new backup scheme with a backup job, a backup copy and a tape job.

My idea is to have a daily backup job with only 1 restore point, and a backup copy job to archive GFS tree of that dailies, and two tape jobs (weekly and monthly), which copy the daily one to tape when necessary.

The problem is that I created a backup job with 1 restore point, forward incremental, with a monthly active full to renew the vbk, but I have launched it three times and I have one vbk and two vibs on disk.... and I was expecting to have only the vbk (which gets increments merged every day)...

What am I doing wrong??? I don't want any vib in my daily schedule. Is it possible to have a job with only 1 restore point??? Why is ignoring that and not merging the just created incremental into the full vbk file?

Regards.
Dave.

[foggy]

David, since you're mentioning merge, looks like you want to use forward incremental forever mode, am I right? This mode is effective only if both active and synthetic fulls are not enabled.

[Dave338]

Hello foggy.

Then, if I understand it well... merging incrementals to full only works with forever forward incremental???
It was my initial idea, but then I thought that it would be good to have a "clean" new full backup once a month (or two...), instead of having a full file with "infinite" increments merged, so it probably will be getting bigger and fragmented over time...
Is there any method to have this implemented?? (maybe with reverse incremental+active monthly full??)

And, with my current setup, what will happen??? It will be making incrementals until next active full and then will delete entire old chain??

Thank you very much
Dave.

[foggy]

Then, if I understand it well... merging incrementals to full only works with forever forward incremental???

Correct.

It was my initial idea, but then I thought that it would be good to have a "clean" new full backup once a month (or two...), instead of having a full file with "infinite" increments merged, so it probably will be getting bigger and fragmented over time...
Is there any method to have this implemented??

You can do periodic active fulls manually, this will not change the backup method.

And, with my current setup, what will happen??? It will be making incrementals until next active full and then will delete entire old chain??

Correct.

[Dave338]

Hello again foggy.

I have some news (good IMO xD )

I've configured the daily job with reverse incremental and only 1 restore point, with the active monthly, and now is doing what I want. I have only the full vbk which each day is upgraded. The reverse file created vbr is deleted inmediatly due to the 1 restore point config, so it keeps only the vbk file.

By the way, I didn't want the job originally configured with reverse incremental to avoid excessive IO on the repository, but this setting also permits having 2 r.p. maintaining one vbr with the vbk, and my tape jobs will only copy the vbk, and that was my goal after having no success previously with syntethised fulls and the "process incrementals" problem (copies also vib's and old merged vbk to tape because it detects it as a new vbk)

Now I have two options. Leaving the job with reverse incremental, active monthly full, all automated... or forever forward with manual active full (less IO but manual)

Maybe it would be a good option to include on future revisions the ability to make active fulls on forever forward mode... in fact I thought from the beggining that it was possible, it makes a lot of sense to me.

Regards.
Dave.

[foggy]

By the way, I didn't want the job originally configured with reverse incremental to avoid excessive IO on the repository, but this setting also permits having 2 r.p. maintaining one vbr with the vbk, and my tape jobs will only copy the vbk, and that was my goal after having no success previously with syntethised fulls and the "process incrementals" problem (copies also vib's and old merged vbk to tape because it detects it as a new vbk)

Yes, in case of reverse incremental job, incremental files are skipped and only VBK is copied to tape during each job run.

I would also like to note that having just a single restore point retention is not generally recommended as it could potentially result in situations, where you find yourself with no valid restore point available. So setting retention to keep at least 2 might be a better idea.

[Dave338]

Thanks again

I update my retention to 2 restore points, and I think leaving it in reverse incremental will be my best option.

Regards.
Dave.

[foggy]

Yep, at least while your storage manages the higher I/O and your VMs are ok with running longer on snapshots.

[Dave338]

Oh yes!! they were previously making direct to tape backups so some stayed an hour or more with snapshot.

Dave.