restore wizard does not offer all restore points on tape

[VCFP]

Hi,

we are running VBR 10 and put full backups on tape weekly, with an expiration time of 2 months.
The last weekly full backup of every month is kept for several years. We had a look at GFS but
its mode of operation as we understood it was not deemed appropriate for us.

I would like to restore a virtual machine which should exist on all the monthly tapes until
December 2020.

When I started the restore wizard ( Hyper-V / Restore from backup / Entire VM restore /
Entire VM restore / Add VM / From backup) and searched for the VM name I expected to get a list
of all the full restore points that are still on tape, ie. from 2020-12, 2020-11, 2020-10, ...
until the first backup of that VM.
This does not happen. I only see some restore points for 2020-03 and 2020-01 (why those dates and
not all the others?) and from 2018-12 / 2018-11 when we started to use VBR.
Something similar happens when I search for a different VM, only with slightly different dates
for the available restore points.

In the "tape infrastructure" section I can see ALL the expected monthly tapes and in
"Properties / Files" for the tapes I can see the .vbk files, so I guess VBR still knows them.

Why is only a small subset of our full backups visible in the restore wizard? Doesn't VBR keep
information about restore points in a database online?
Is this database not a full catalog of every restore point on all the tapes? If not, what is the
point of this database?

I would expect to find all restore points of a virtual machine still on not-expired and
not-freed tapes offered for a restore.


Regards,

Klaus

[Dima P.]

Hello Klaus,

Restore options are different for backup files stored on tape; for example you can restore vm from the backup on tape directly to production but you cannot start file level recovery from vm backup on tape. Currently, you can find all the backups stored on tape under the Home > Backups > Tape node and launch a specific restore from tape option. I'll note an improvement request based on your post, thanks for the feedback!

[VCFP]

Hi Dima,

thanks for the information about the "Tape" node. I had never used that before and found all the restore points I was missing.

BUT...

I think the user interface could use some improvement, both on that "Tape" node and in the restore wizard.
For us the tape node shows something like this:

Code: Select all

  the-jobname on Tape    2019-01-26...    MONTHLY-pool (duplicated)
  the-jobname on Tape    2020-02-29...    MONTHLY-pool (duplicated)
  the-jobname on Tape    2019-01-26...    MONTHLY-pool
  the-jobname on Tape    2020-02-29...    MONTHLY-pool
  the-jobname on Tape    2021-02-06...    WEEKLY-pool (duplicated)
  the-jobname on Tape    2021-02-13...    WEEKLY-pool (duplicated)
  the-jobname on Tape    2021-02-20...    WEEKLY-pool (duplicated)
  the-jobname on Tape    2021-02-06...    WEEKLY-pool
  the-jobname on Tape    2021-02-13...    WEEKLY-pool
  the-jobname on Tape    2021-02-20...    WEEKLY-pool

Some of those entries contain many restore points and some only two or three.

Why are there multiple entries for a specific job name (always the same in the example above) which always goes
to the same tape pool (say, WEEKLY-pool: 3 entries)?
My guess is, some internal identifier (maybe for the pool or the jobname) changed due to some action by us.
But people (or at least I myself) always use names (pool names, VM names) to navigate and search and not
internal identifiers (which are not even visible above and thus cannot explain why there are multiple entries
for the same thing).

What I would expect from a software like Veeam is using its internal identifiers (of course unique keys are
necessary in a database ) but giving the user a consolidated view (of that database) made for people.

Code: Select all

  the-jobname on Tape                     MONTHLY-pool (duplicated)
  the-jobname on Tape                     MONTHLY-pool
  the-jobname on Tape                     WEEKLY-pool (duplicated)
  the-jobname on Tape                     WEEKLY-pool

Each entry should contain all the restore points for a VM (possibly multiple entries for the same VM). Then
I can select one of them and start "restore entire VM" and choose the specific restore point I want.

Regards,

Klaus

[Dima P.]

Klaus,

Thanks for the heads up! Mind me asking if the tapes with duplicated backups were ever marked as free or removed from Veeam B&R configuration?

[VCFP]

The string "(duplicated)" is part of the name of the media pool.
We have 4 pools "WEEKLY-pool", "WEEKLY-pool (duplicated)", "MONTHLY-pool", "MONTHLY-pool (duplicated)"
in order to get two tape media sets for the backup job. The media sets are then stored in different locations.

Weekly/monthly full backups are copied to tape by two tape jobs "disk-to-tape" and "disk-to-tape (duplicated)", chained
together. This is why I would expect 4 entries (one backup job to 4 different media pools) and not multiple entries
for the same pair of job name and media pool.

I suppose tapes may have been marked as free or have been moved between pools. Is this relevant for the question
why we see so many entries under the "Backups / Tape" node in VBR?

[Dima P.]

VCFP,

Understood, thanks! Since these are different tape jobs currently backup files are displayed separately per-job. I'll note an improvement request based on your post and will discuss it with RnD folks. Cheers!