GFS tape writing to yearly instead of monthly media set?

[pkelly_sts]

[Case #02052043]
I have two GFS Tape jobs (1 per site) which are each configured to only run Monthly copy (on last Friday of month) and yearly copy (on last Friday of year) however one of them this morning is attempting to write to the yearly media set.

Any ideas why that would be? I've opened a case but in case support are busy I thought I'd ask here as I'd really like to get this one resolved in time to run todays copies correctly.

[lyapkost]

Hi, Paul. The job started today at midnight to copy backups to the Monthly media set according to the schedule and waited for them until 9 AM when you've terminated it. At the same time after start the job checked past gfs dates and for some reason didn't find backups from 2016-12-30 (Yearly media set) in the database and decided to rebackup them.

[pkelly_sts]

Thanks Konstantin, I got a swift response from support stating the same (I hadn't spotted the yearly fail).

If I "Start" it now, will it restart again? What will happen with the backup of that machine that would normally run tonight, and get copied to GFS tonight, as the monthly?

[lyapkost]

If started manually the job of course will restart, and will do absolutely the same: continue waiting for source job to perform backup. As soon as it appears in repository, tape job will copy full (or synthetic full) backup to the Monthly media set. And it will also try again to copy yearly backup from 2016-12-30 to tape. In case you want the tape job to 'forget' about previous gfs intervals you need to disable and enable the tape job. If done so, the tape job will work only with backups which appeared after the moment it was enabled and won't search back for skipped gfs points for past dates.

[pkelly_sts]

Excellent, useful tip for "resetting the clock" by disable, thanks, will start the job again...

[pkelly_sts]

Veeam is a fantastic tool but there are days when it drives me insane with its "logic" !!!

For whatever reason, it didn't want to continue the yearly backup to the yearly media set so I added a "blank" (previously used in BExec) tape on Friday.
It didn't like this tape for some reason, producing the following cryptic error:

29/01/2017 19:02:16 :: Library locations are out of sync, please run the inventory. Tried to load tape IO4105L6 into the drive, but IO4105L6 was loaded instead.

So, even though the error made no sense I erased the tape anyway and have re-run the job.

Now the job has decided to use THIS tape as the continuing monthly media set (which is set to append, and which has the last tape still available with 1.9Tb free!) so when it gets to the outstanding yearly data (next in the list) it won't have a fresh tape for it to write to the Yearly media set, again, and with this being a remote site I cannot get another blank tape to it!!!

I really wish it would have just carried on in the first place with the MONTHLY tasks that it COULD complete and queued the Yearly task to the end so on a retry, only the yearly task needs to be completed?

Frustrated!!!

[lyapkost]

Paul, please keep in touch with support team to sort things out. There is an option to place a tape in GFS media pool to the specific media set to avoid situations like you've described: Media Pools - edit your GFS media pool - 'GFS Media Set' step - Advanced - select a media set and add tapes from the media pool there, these tapes won't be used in other media sets.

[pkelly_sts]

Hi, yes, I'm still in touch with them, unfortunately a single reply a day makes it a very laborious process to resolve, hence why i thought I'd post here as well.

I understand that you can add media directly to pools as you describe, but in this case I believe it shouldn't actually be necessary...