How to realize a GFS backup on repository and tapes

[PGA]

Hi to everyone.
I’am starting to use Veeam Backup & Replication 8.0 and this is my first experience with Veeam, so please excuse me if my questions will be stupid. 
I have to implement the backup of my servers using this GFS schema:
Mon Tue Wed Thu Fri
Day1 Day2 Day3 Day4 Week1
Day5 Day6 Day7 Day8 Week2
Day1 Day2 Day3 Day4 Week3
Day5 Day6 Day7 Day8 Month1
Day1 Day2 Day3 Day4 Week4
Day5 Day6 Day7 Day8 Week5
Day1 Day2 Day3 Day4 Week6
Day5 Day6 Day7 Day8 Month2
Day1 Day2 Day3 Day4 Week1
Day5 Day6 Day7 Day8 Week2
Day1 Day2 Day3 Day4 Week3
Day5 Day6 Day7 Day8 Month3
Day1 Day2 Day3 Day4 Week4
Day5 Day6 Day7 Day8 Week5
Day1 Day2 Day3 Day4 Week6
Day5 Day6 Day7 Day8 Month4
Day1 Day2 Day3 Day4 Week1
Day5 Day6 Day7 Day8 Week2
Day1 Day2 Day3 Day4 Week3
Day5 Day6 Day7 Day8 Month5
Day1 Day2 Day3 Day4 Week4
Day5 Day6 Day7 Day8 Week5
Day1 Day2 Day3 Day4 Week6
Day5 Day6 Day7 Day8 Month6

Where the Daily backups are Incremental (retention =8) and the Weekly and Monthly backups are Full (retention = 6 for Weekly and Monthly too).
I want to save the backups on the repository and also to copy them on tapes, using the same retention described above.
Regarding the tape, I have a standalone drive (HP Ultrium 4) without media charger, so I have to insert manually the tape in the drive and for this reason I need to schedule the weekly and monthly full backups on a working day (Friday).
I have already read the User Guide of Veeam, but unfortunately it is not so clear to me. 
From the guide, I understand that to realize a GFS backup it is necessary to use a copy job, because only in the copy job there is the possibility to define the retention for Weekly, Montlhy, etc. backups.

So, to implement my GFS schema, as first step, I have to create a “Forward Incremental Backup” job from Monday to Friday with an “Active full Backup” scheduled on Friday. The restore points of this job is 8. This job will use the repository as destination.

Then, as second step, I have to create a “Backup Copy” job for the Weekly and Monthly backups. In this copy job I can define the retention for the Weekly and Monthly backups.
This copy will use the same repository of the previous backup job.

Now the questions:
1) Do I correctly comprehend how Veeam implements the GFS backup?

2) The copy job has got a retention for the Weekly, one for the Monthly, one for the Quarterly and one for the Yearly backups. The meaning of these
is clear for me.

I do not understand the meaning of the “Restore points to keep” value.
Is it the sum of the others four restore points? (Restore points to keep = Weekly + Monthly + Quarterly + Yearly).
Can you explain me the meaning?

3) The copy job runs one per day, it will copy also the daily restore point? The incremental ones?
If it is so, which is the retention used?

4) How can I copy the GFS backups on the tapes?

Help me please
Thanks in advance for the time you will spend for me!
Ciao

[Shestakov]

Hello Fabio and welcome to the forums!

1) Yes you do!

2) No, it`s not the sum of restore points, but the number of daily restore points. Weekly and Monthly restore points will be kept in accordance to their own policy.

3) Backup copy job copies the latest VM restore point from source to destination. In your case from Monday to Thursday incremental backups will be copied, while on Fridays it will be full backups named in accordance with the policy.
Note that one and the same full backup can be marked as weekly and monthly. When transforming weekly, monthly chains, Veeam Backup & Replication checks the flags set for the full backup file. If the full backup file belongs to some other retention policy tier and must be retained on the target backup repository, such backup file will not be removed.

4) Do you mean only weekly and monthly ones?
If yes, easiest way will be to do next:
Create a media pool called “Weekly” with retention period equal to 6 weeks.
Create a backup to tape job called “Weekly” and point to the “Weekly” media pool and schedule it to run on weekly basis.
Create a media pool called “Monthly” with retention period equal to 6 months.
Create a backup to tape job called “Monthly” and point to the “Monthly” media pool and schedule it to run on monthly basis.
In the next versions we are planning to make GFS for tapes more intelligent.

Thanks!

[PGA]

Hi Nikita,

first of all, thanks a lot for your fast reply!

Now it is more clear to me.

I have some additional questions:

a) In my case, the repository used as destination of the copy job is also the source repository, that is the repository used by the backup job.

Is this a problem? Is it preferable to use two different repositories?

If it is not a problem, then source and destination repository can be the same, when the copy job works, does it create a copy of the original backup files, and apply the GFS retention to these copy, or it simply updates the "retention flag", that you have explained me, to the original files, to apply the GFS retention directly on those?

b) Regarding the tapes, I need the tapes of all the backups, so daily (incremental from Monday to Thursday), weekly (full on friday) and monthly (full on Friday).

How can I do?

Thanks again for your precious help

[Shestakov]

Fabio,

a) Technically you can create 2 backup repositories on the physical repository: one as a source, another as destination, but I don`t see a value of pointing the backup copy job to the same repository.
The main purpose of Backup copy is making another copy of VMs` backups for the better data protection reliability. We basically recommend to follow 3-2-1 backup rule and backup copying allows you to create several instances of the same backup file in different locations, whether onsite or offsite.
Could you please explain your use case?

b) It`s feasible as well. You can easily customize it by editing backup to tape job called "Weekly" (referring to my previous post). On the step Media Pool check "Process incremental backup files" and point a different media pool for incremental backups if needed. Schedule the job to run Mon-Fri with full backup on Fridays and it will work.

Thanks!

[PGA]

Hi Nikita,
regarding your question: Could you please explain your use case?
I have understood that a copy job is the only way to have a backup schema that follows the GFS retention I described in my initial post.
It seems strange also to me that to do that I have to use a copy job (because in my case it is not a real copy), but I do not see another way.

How could I do?

Thanks

Ciao

[Shestakov]

Thanks for the clarification, Fabio. I see your point now.
The initial idea of having the protected environment is to make most recent backups on the onsite backup repository and another copy offsite, including those historical copies intended for long-term archiving(GFS). That`s why GFS is available only for backup copy jobs for now.

For your situation, if you have just one repository and want to implement GFS retention I would suggest to decrease number of "Restore points to keep" in the backup copy job to 2(minimal available value), since you will have exactly same restore points from your source backup job.
Or just simply schedule several basic backup jobs to make daily, weekly, monthly backups.

I`ll take into consideration that GFS retention can be also usefull for backup jobs for some cases. Thanks for the feedback!

[PGA]

Hi Nikita,
yes I think that a function to obtain directly a GFS backup could be a new feature for Veeam.

In the meantime I will try to implement the GFS backup with several backup jobs, as you suggested.

To do that, I need another information

Scenario:
I scheduled an incremental backup job from Monday to Thursday.
I have another separated backup job scheduled to have a Full backup only on Friday.

On the first week I will obtain these results:
Mon --> Full_backup_1 (the first execution of the incremental backup job).
Tue --> Incremental_1 (differences from the Full_backup_1)
Wed --> Incremental_2 (differences from the Incremental_1)
Thu --> Incremental_3 (differences from the Incremental_2)
Fri --> Full_backup_2

On the second week I will obtain these results:
Mon --> Incremental_4 (differences from the incremental_3 or from the Full_backup_2? ).
Tue --> Incremental_5 (differences from the Incremental_4)
Wed --> Incremental_6 (differences from the Incremental_5)
Thu --> Incremental_7(differences from the Incremental_6)
Fri --> Full_backup_3

The question is: Does Incremental_4 contain the differences from the incremental_3 or from the Full_backup_2?

In another words, do the two backup jobs work on different layers or they share the information about the backup status of the files?

Thank a lot

Ciao

[Shestakov]

If you schedule several backup jobs, they don`t collaborate between each other, so if you want to have 8 daily restore points for the last 8 days, 6 weekly and 6 monthly restore points, I would suggest to schedule:
1st backup job with 8 restore points and make it run daily on weekdays. It will have 1 Full + 7 increments
2nd job with 6 restore points to run weekly(on Fridays). You can make it have its own chain of 1 Full + 5 Increments or just make 6 Fulls.
3rd job with 6 restore points to run monthly(on the last Friday of the month). You can also make it have its own chain of 1 Full + 5 Increments or just make 6 Fulls.

Let`s say you decided to make only full backups for 2nd and 3rd jobs, so after 6 months you will have 13 Full backups and 7 Incrementals on your backup repository. And once a new backup file appear, another(oldest in its job) will be removed in accordance with the retention policy. Hope that helps. Thanks.

[PGA]

Hi Nikita,

thanks a lot for your precious help!

Ciao

[Shestakov]

You are welcome, Fabio!
Glad that you found needed answers.

[foggy]

3) The copy job runs one per day, it will copy also the daily restore point? The incremental ones?
If it is so, which is the retention used?

Backup copy job does not simply copy backup files created by the corresponding backup job, but rather synthetically creates restore points on target repository using data available in the source repository, copying the latest VM state during each synchronization interval. I recommend reviewing the corresponding user guide section to get better understanding on how backup copy job retention works.