Understanding Backup Copy Jobs and GFS

[andreash]

Hi,

fairly new to Veeam and going through a steep lerning curve, I'm afraid.

I have inherited a Veeam 7 Installation that backs up several (~10) VMs to an iSCSI attached NAS using a set of three schedules:

1. daily backups, keeping 14 restore points. This is to provide daily backups for two weeks. It is configured with Standard Incremental and Synthentic Full, so I have more than 14 restore points on disk, dependent on the oldest available backup. I guess moving to Reverse Incremental would make sense here.
2. weekly backups, run every saturday, with 4 restore points. This actually creates _two_ backups every week, one incremental saturday night, one full sunday.
3. monthly backups, run every last sunday of the month. This creates full backups _every week_.
All three backup jobs back up into the same repository.

Being puzzled why the weeklies create two backups, and the monthlies run weekly, I have looked into Backup Copy Jobs to provide my GFS retention, but I'm not quite sure how to set it up.

Do I Keep my daily schedule, and configure one Backup Copy Job to run daily, keep two restore points, and keep 4 weeklys and 5 monthlys for archival purposes
or
Do I only keep two (or even one?) restore Points in my daily schedule, and configure the backup copy job to run daily, keep 14 restore Points, and keep 4 weeklys and 5 monthlys for archival purposes

The second option looks the most straight Forward to me, having all backup/archival settings in one job, and probably taking most advantage of VEEAM dedup. No backup target dedup available, I'm afraid.

Any help/directions are greatly appreciated.

Regards,
Andreas

[Vitaliy S.]

Hi Andreas,

You need to keep your daily schedule of the regular backup job and configure your backup copy with the GFS settings you want: keep X number of weekly backups and X number of monthly backups. Each restore point created according to the GFS settings will be a full backup.

Thanks!

[foggy]

2. weekly backups, run every saturday, with 4 restore points. This actually creates _two_ backups every week, one incremental saturday night, one full sunday.
3. monthly backups, run every last sunday of the month. This creates full backups _every week_.

Most likely you have synthetic full backup enabled on the specified day, aside from periodic job schedule.

[andreash]

You need to keep your daily schedule of the regular backup job and configure your backup copy with the GFS settings you want: keep X number of weekly backups and X number of monthly backups. Each restore point created according to the GFS settings will be a full backup.

Hi Vitaly,

thanks for your quick feedback. I have changed now to:
- Backup Job "Daily Backups", Reverse Incrementals, 14 Restore Points
- Backup Copy Job "GFS", copy every day, keep two restore points, and keep 4 weekly/6 monthly restore points for archive.

Is there anything else I could do to optimize my setting?

Most likely you have synthetic full backup enabled on the specified day, aside from periodic job schedule.

Spot on! Thanks!

[foggy]

Is there anything else I could do to optimize my setting?

Not really, as it's a pretty straightforward setup. Btw, weren't you considering to have offsite copies of your backups also?

[andreash]

Not really, as it's a pretty straightforward setup. Btw, weren't you considering to have offsite copies of your backups also?

Not for now. Offsites will come later, and will simply be another copy job to a different repository.

Thanks a lot for your quick help!

[andreash]

Ok, looks like I still get something wrong here.

For simplicity I had decided to have both my daily backup job and the backup copy job use the same target repository. This was meant to have all daily, weekly and monthly backups in the same location.
Unfortunately the backup copy job does not copy data. It simply sits there waiting for new restore points, even when the daily backups have finished.

So, does the backup copy job automagically use existing backups in the repository, and keep track of the GFS settings, or do I have to set an Independent target repository (since moving from an existing setup this would mean to resize nearly all of my backup repositories)?

[foggy]

This behavior is expected, since it is by design for backup copy job not to look for source restore points in the repository that is specified as the target.

[veremin]

If possible, you can create additional folder on the target device, add it as a repository to backup console and specify it as a target for backup copy job. This way, the above mentioned limitation will be bypassed.

Thanks.

[andreash]

A late follow up question after having set this up and working fine since several months, I still have the feeling things could be optimized on the backup copy Job side.

Current Setup is like this:
Regular Backup: Reverse Incremental, 21 restore points to keep. I will add synthetic fulls every 4 weeks, other than that I'm happy with this setup
Backup Copy Job: Run daily, keep 2 restore Points (Minimum possible), GFS Retention with 4 weeklies every sunday and 6 monthlies every last day of the month.

Basically, the Backup Copy Job is only there to populate the GFS scheme.

Do I really have to run the Job daily, or is it sufficient to run it every saturday or sunday, and save the Backup Transformations every day?
I'm worried what happens if the last day of the month is a monday, and my Backup Copy Job only runs on weekends. Would it still create a Restore Point for the correct day?

Kind Regards,
Andreas

[foggy]

Do I really have to run the Job daily, or is it sufficient to run it every saturday or sunday, and save the Backup Transformations every day?

For pure GFS purpose, it is sufficient to run the job weekly. However keep in mind, that every GFS restore point is a full backup available on this day (not the newly copied VMs state, which is saved in the incremental restore point created at that day).

I'm worried what happens if the last day of the month is a monday, and my Backup Copy Job only runs on weekends. Would it still create a Restore Point for the correct day?

In this case the full backup from the next synchronization interval will be used. The tricky restore points selection algorithm is described in the corresponding user guide section, worth reviewing.

[andreash]

Hi foggy,

thanks for your clarification. I had reviewed the user guide, but with english not being my native language I wasn't quite sure I understood everything correctly.


Kind Regards,
Andreas

[foggy]

Glad I could help. Let us know if you need any further explanations.

[claudiofolu]

Sorry for the late followup on this, but I dont undestand this statement:

This behavior is expected, since it is by design for backup copy job not to look for source restore points in the repository that is specified as the target.

Why is that? I also have my backup copy job pointed to the same repository and is the backup copy job is not copying anything.

[foggy]

There is an inbuilt mechanism that prevents backup copy job from looking restore points in the repository that is specified as backup copy job target (to avoid copying restore points created by itself). If you need to store both backup and backup copy files in one location, you can simply create an additional folder there and specify it as a separate target repository for backup copy job.