Getting insert a tape to continue

Based on your configuration it should work fine.  I might suggest opening a support ticket for better troubleshooting.

Can you check the job configuration to see if the option to use a specific number of simultaneous drives is enabled? Perhaps the AWS Tape Gateway is simulating physical tapes but with a different behavior. Some settings may cause Veeam to interpret that a tape has been "ejected" and require a new one. Check the AWS Tape Gateway settings to see if there is any automatic tape locking or expiration policy in place.

AWS says I can run up to 10 per their performance recommendation https://docs.aws.amazon.com/storagegateway/latest/tgw/Performance.html#performance-tgw

(Your backup software can back up data on up to 10 virtual tape drives on a Tape Gateway at the same time.)

I have 8 tape jobs, half of them failed while the other seems okay. With this fail rate, I’m not confident if they actually got the weekly backup. B&R seems to struggle with the GFS weekly while the daily seems OK.

All 8 of them works fine for daily, but they fail on the weekly, and they failed on the monthly when we went from Feb to March.

AWS do have tape retention locking, but I’m not using that feature as I’m just using a regular tape pool not custom one (https://docs.aws.amazon.com/storagegateway/latest/tgw/CreatingCustomTapePool.html)

B&R seems to see the tape drive but not detecting the media, it’ll say drive is in use by the job’s name but state is empty.

There is this setting of the type of media changer. I wonder if B&R is incompatible with it, I think I’m going to change the type before I submit ticket. I’m using the AWS-Gateway-VTL media changer type.

Try the IBM version in that screenshot and see if that works.  It could be the media changer type and this is a good test.

i created a new tape library because I couldn’t switch it. Let’s see if this works tonight and this weekend.

Sounds like a plan.  Let us know how it goes.

Hi ​@perkedel

I don’t think this is about the emulated hardware, likely it’s about media sets. If the media set was closed for some reason (check bottom of the page, sometimes it may close out the media set earlier than configured), then the tapes would not be writable until the tape protection (retention) period is complete.

Can you show what you have configured here: https://helpcenter.veeam.com/docs/backup/vsphere/backup_to_tape_gfs_options.html?ver=120

With AWS VTLs, the export tape command tells the VTL to archive the tape, and while I see they’re all online in your screenshot, I would still confirm this setting.

Basically, I suspect the media set closed, but this is best reviewed with Veeam Support. The logs should tell pretty clearly why available tapes weren’t selectable.

If you want to do some research yourself, tape logs are pretty human readable; in your case, you need to magic search terms in the Tape GFS Job Log:

will be placed

mediapool

Since it’s Tape GFS and we might need to put backups into different GFS media sets, searching will be placed will give you a list of each candidate found to backup as well as which media set it’s going to -- often you can catch the tape job “backfilling” missed GFS periods, for example, it might be a weekly GFS day, but the last monthly GFS point was missed due to hardware error previously -- Veeam will try to fill this missed GFS period on future runs, and thus you might get a request for tapes from media sets you weren’t expecting.

mediapool is a keyword you can use to jump straight to tape selection. From the start of the run where you got the unexpected request for a tape, search downwards for mediapool until you find an XML about the mediapool. Start reading line by line there as this is where tape selection starts, and it should tell in plain language which tapes were available, and why they were or were not chosen.

Thanks for that link, I verify that the check box is not on. I believe by default the gfs wizard does not have this turned on. Here’s what’s on all of them, no check on the Export the following media sets upon job completion.

I’ll take a look at the log, thanks for the pointer for the mediapool keyword! Makes it easy to pin point.