backup - Copy - Tape Strategy question

[rmueller]

I am a little confused as to how to accomplish a tiered backup strategy (local, remote copy, remote tape) that would allow me to recover any file created on a given server within 365 days of its creation...

As an example: I have a backup for a VM that runs daily at 8PM, it is set to do forever incrementals, maintain 31 days of recovery (locally); synthetic fulls on Saturday.

The VM's backup copy job takes the local data and moves it to a second remote machine/storage location. That remote machine has Veeam running and a tape library.

This VM's backup copy job is scheduled at at 10PM it is set to retain 7 restore points.

Ideally, I'd like the 31days of local restore points, then some number of remote copy restore points and a mechanism whereby I can rollup a weekly full and copy it to tape.

The confusion is around the archival settings...

If I only set the archival for 2 - weekly, and once a week backup the oldest weekly - does that get me what I want? I assume a setting of 2 keeps one from the prior week, creates a new for the current week, and on the third week deletes the oldest; is that correct?

ron

[veremin]

Hi, Ron,

If you want to copy full backup to tape on weekly basis, I don't think you should bother with GFS archival scheme. You can leave everything as is, create file to tape job, select backup copy job folder as a source for it, specify corresponding file mask (.vbk) and set this job to run on weekly basis. This way, file to tape job will archive .vbk file produced by backup copy job on weekly basis. Keep in mind that tith 7 restore points, the .vbk file will be 7 days old.

Thanks.

[rmueller]

I had that working a few weeks ago, it does accomplish getting a weekly copy to tape. But...

What I want to graduate to is a tiered backup solution to maximize my on-disk restore window within both a local and remote site, with Local, Remote and Remote Tape in the mix. Something like this...

Local storage - as many calendar weeks as I can store, with synth fulls once a week.

Remote - one recent full + associated incrementals (the number is equal to the number of restore points) and a weekly archive that gets created on for instance on Sunday

I'd like to have a one or two week overlap between Local and Remote (same restore points in both places).

Then each Monday, I'll run a file-to-tape job (and a little powershell script to hardlink the oldest vbk files to a separate directory) to get the oldest vbk file (which should be the weekly archive file) on to tape.

THREE things I'd like to know:

1 - is there any calculator or tool that will show me what will end up on disk; given a particular configuration of backup/backup-copy/backup-copy archive settings
2 - does the archive (ex. keep 1 weekly archive copy) actually get created every week on the specified schedule, or only after the number of specified restore points have been reached
3 - what settings in the backup job, backup-copy job would I need to use to get the following 52 week setup:
Last 25 weeks local - incr with weekly fulls
Weeks 30 - 24 on the remote copy site (weeks 24 and 25 overlap with the local storage)
Weeks 31 - 52 on tape as a weekly rollup

assuming the following
175 local restore points (25 weeks - combination of fulls and daily incr)
Last two weeks of the local storage overlapping on the remote so week 24 and 25 on the remote as well as the local
I want 6 weeks of on disk restore points: so 42 Daily restore points on the remote (starting at week 25 running back 4 additional weeks)
1 weekly archive on disk (the once a week rollup of the last week of the 6 weeks of on-disk restore points)
Weekly rollups going back 20 more weeks on Tape

I know how to set the backup job - but how do you set the backup-copy job to accomplish the above...
BackupJob - 175 restore points with Synthetic fulls each Sunday (I know this one - works perfectly); I have set 175 restore points, Incremental, enabled synth fulls and specified Sunday; and the secondary target is set to the backup copy job

ron

[veremin]

1 - is there any calculator or tool that will show me what will end up on disk; given a particular configuration of backup/backup-copy/backup-copy archive settings

No, there is no such a tool that makes such an estimation.

2 - does the archive (ex. keep 1 weekly archive copy) actually get created every week on the specified schedule, or only after the number of specified restore points have been reached

When the specified retention is reached, the full backup start moving forward on regular basis ("copy interval"). When it reaches day specified for GFS archival, the new GFS restore point is automatically created.

3 - what settings in the backup job, backup-copy job would I need to use to get the following 52 week setup:

I doubt it will be possible. Backup copy job creates one restore point per copy interval. So to say, if copy interval is equal to 1 week, the backup copy job will create 1 weekly worth restore point. You can't make it copy the 24-30 week old restore points. Neither can you specify a delay for it, so that, it processes only old restore points.

I had that working a few weeks ago, it does accomplish getting a weekly copy to tape. But...

You can stick to the proposed plan or utilize Robocopy (or similar tool). In the latter case, you should write custom script that instructs Robocopy to copy to target location files older than 24 week. Then, you should set this script as post job activity.

Thanks.

[foggy]

1 - is there any calculator or tool that will show me what will end up on disk; given a particular configuration of backup/backup-copy/backup-copy archive settings

There's no such a tool, however you could perform a rough estimate of the space required.

[rmueller]

OK...just so I totally understand what you are saying...

"When the specified retention is reached, the full backup start moving forward on regular basis ("copy interval"). When it reaches day specified for GFS archival, the new GFS restore point is automatically created."

I read this as the most recent full gets created after the retention count is reached - agreed. But the second part, says to me that the archive is created when the day specified is reached...and doesn’t say that the archives are dependent on the retention count or are they?


Then the whole thing about turning off the GFS and make a separate full, how does that make a difference, what does that even mean?

Scenario:
Last 25 weeks local - incr with weekly fulls Weeks 30 - 24 on the remote copy site (weeks 24 and 25 overlap with the local storage) Weeks 31 - 52 on tape as a weekly rollup
assuming the following
175 local restore points (25 weeks - combination of fulls and daily incr) Last two weeks of the local storage overlapping on the remote so week 24 and 25 on the remote as well as the local I want 6 weeks of on disk restore points: so 42 Daily restore points on the remote (starting at week 25 running back 4 additional weeks)
1 weekly archive on disk (the once a week rollup of the last week of the 6 weeks of on-disk restore points) Weekly rollups going back 20 more weeks on Tape

DESIRED
So with dots... it looks like this
Past 1-year mark |----------------------------------| current time
................... -- imagine this to be 175 restore points with combo of weekly fulls/Incr on the local backup site
WR............... -- imagine this to be the remote site holding the remainder of the year with a two week overlap
...................... -- imagine this to be tape storage of weekly fulls - going back as far as possible

COMPROMISE
So with dots again...this is what I had in mind instead. Keeping, one full (rollup of incr restore points), and one archive full (rollup of most recent full + incr restore points)
Past 1-year mark |----------------------------------| current time
................... -- imagine this to be 175 restore points with combo of weekly fulls/Incr on the local backup site
WR................................. -- imagine this to be the remote site holding as many incr's as possible with two fulls (one full created weekly)
................................... -- imagine this to be tape storage of weekly fulls - going back as far as possible

REALITY
But I think I can only get this out of the product

Past 1-year mark |----------------------------------| current time
....................... -- imagine this to be 175 restore points with combo of weekly fulls/Incr on the local backup site
WR..... -- remote site holding 7 restore points and a weekly archive
................................................ -- tape storage of weekly fulls - going back as far as possible, using a script to link the oldest vbk file and do a file backup to tape each week

If I'm stuck with the last option (REALITY) so be it, but still don't understand why (COMPROMISE) doesn't work...lastly where would something like DESIRED (a mechanism to accomplish a tiered backup solution) rank on a new feature list for this product?

ron

[veremin]

I read this as the most recent full gets created after the retention count is reached - agreed. But the second part, says to me that the archive is created when the day specified is reached...and doesn’t say that the archives are dependent on the retention count or are they? Then the whole thing about turning off the GFS and make a separate full, how does that make a difference, what does that even mean?

Let me give some example. Probably, this will clarify things for you.

Imagine that the retention count is reached and full backup (.vbk) starts moving forward. The current day is Monday, and Wednesday is specified for weekly GFS point. What happens next is:

1) Within next two days the full backup will reach Wednesday, the day specified for weekly GFS point. So to say, on Wednesday the oldest restore point will be presented by full backup (.vbk).
2) Once it happens, the VB&R will create new independent full backup, using the oldest restore point (.vbk). This backup will be a GFS weekly point. It will be preserved in accordance with GFS retention scheme.

Thanks.