GFS Active Full

[ekisner]

Hi there.

With regards to GFS jobs, I'm finding that on occasion it misses some of the disk jobs. For example, just now, it failed because it started whilst a disk job was running, and in the GFS details it shows that the tape job for that particular backup has already failed - reason being "Source job retry is required. Canceling..."

When I retry that job, however, I have no way to tell the GFS job to do an active full weekly backup just for that job.

Is there a work-around? Some way to tell it to do an active full for just certain disk jobs?

[Shestakov]

Hi Erik,
there is no such a thing as "active full for just certain disk jobs". It makes sense to make an active full GFS from the scratch rather than having 2 halves of a GFS backup.
The job retries several times to copy all the restore points, if some of them are unavailable and the job is failed, you should make a new GFS backup manually.
Thanks!

[ekisner]

Well that would work if not for the fact that half of my library is now write protected. Sure I can erase the tapes and start the job again, but in no way does it make sense to back up backups a second time when they are already done.

I am home now but tomorrow I will post the job report. A disk job cancelled the tape job. Unless I am to pause my disk backups for 2 days, I cannot do a full take backup.

[veremin]

Not sure whether you find it an applicable workaround, but you can try to remove temporarily unneeded backup jobs from the source of GFS backup to tape job, execute GFS now option, and re-add the previous jobs. Thanks.

[ekisner]

Hmm... definitely a work-around but it gives me an idea (more of an extension of your suggestion). I can create 2-3 different tape jobs, targeting the same GFS media pool, correct? This way, there's much more flexibility.

Since it's relevant, I'm going to hopefully continue this thread. I can open a support ticket if there's a recommendation for such, but in the interim this will be a good source of information for said ticket if it reaches that point. I tried to paste the HTML report, however the formatting was lost making it hard to read. Please accept the following summary, as it relates to a specific disk backup which has no recovery point on tape. It is a job specific to all of my file servers (a particularly important one!).

April 19 :: 4:15pm
- This was a manually triggered "Start" job, not "Active Full"
- All covered disk jobs successful, no data backed up.
- File server job succeeded with 0.0KB backed up

April 19 :: 10:06am
- This was an "Active Full - Weekly" job
- 3 disk->tape jobs succeeded, 1 disk->tape job failed due to error, 13 disk->tape jobs were cancelled due to one of the underlying jobs starting
- File server job failed due to backup cycle being cancelled

April 18 :: 10:00am
- This was a manually triggered "Start" job, not "Active Full" and not schedule
- Many disk->tape jobs succeeded. None failed, however with exception to Endpoint jobs, there were no full backups
- File server job succeeded with 0.0KB backed up

April 16 :: 12am
- This was an automatic trigger, due to job settings telling it to do a full backup
- Several jobs succeeded (actually succeeded), several finished with warnings, a number of them failed because "full backup cycle was cancelled"
- File server job succeeded with 0.0KB backed up


So far to this point, the GFS job is incorrect in its assertion that there is a full tape backup - the last full backup that I have on tape for that file server job is my end-of-month tape job from March 26, where there should be one every Saturday.