Offsite backup

[bostjanQ]

Hi.
We are using Veeam agent free version and Veeam Backup and replication community edition.
There are 4 servers in the environment.
1 Hyperv
2 virtual machines
and 1 physical server.

We added 5th server in our environment which is a server located in cloud. The purpouse of that server is to have backup copies stored there in case of catastrophic event.
We have a problem with understanding best practice of retention policy and how to set them right.
We have bandwith limitations and because of that we must reconfigure our backup policy to meet those bandwith limitations.

Ideally we would like to setup our backups:
- 1st of the month FULL BACKUP, and all the rest of the days in that month to be incremental
- We would be daily syncing those backups to offsite location
- In next month when the new full backup would be created we would like to sync that full backup and after the sync is done a full backup (and all the incrementals from the last month) should be deleted on both sides

QUESTIONS:
Can you please suggest us if that is possible to achieve?
How should we set retention policy, active backups in Veeam B&R and in Veeam Agent?
Question regarding incremental backups: If you have 1FULL backup from the 1st day of the month and then incrementals daily, what if you don't have all the incremental backups stored on offsite location, can you still make a restore or do you need all of the incremental backups to be able to do a restore?
-How can you setup retention policy if one month has 31 days, the other month 30, etc?

Thank you
With best regards

[Natalia Lupacheva]

Hi Bostjan,

1. Could you please give us a bit more details about this 5th server?
Do you have a network connection between Backup server and this server? If yes, backup copy job should help.

2. You can configure periodical active full backups on job settings. Retention policy is configured on Backup storage settings step.

3. Even if you have a corrupted incremental backups chain, you can always restore from Active full backup.

4. You can find the detailed description of the retention policy for Active full backups in User guide.The "old" active full backup and backup chain will not be removed until you've got the new one, it is not depending on the number of days in current month.

Thanks!

[bostjanQ]

Hi Natalia.
First of all, thank you for your reply.

1 - We have a network connection between Backup server and this server. On the server in cloud there is nothing installed (yet), it's a "blank" Windows Server 2019 desktop version.
We planned to sync Veeam backups from primary site to this server in cloud with SFTP server + (Freefilesync.org 3rd party application).
Should we instead of using SFTP+3rd party application to sync backups from SITE1 to SITE2 rather install Veeam Backup and Replication on server in Cloud, create a repository there and then schedule backup copy job on Backup Server in SITE1 to a "new repository" on Cloud server?

2. What should the retention had to be set for our specific scenario where we would like to have 1 FULL backup and incremental backups for the rest of the days in that month?
30? What if the month has 31 or 28 days?
We would also like that the first (active) full backup in the new month would "clean" all previous full and incremental backups from previous month if that is achiavable?

3. I didn't quite understand corrupted incremental logic. Please, let me describe with an example.
If we would have 1 FULL backup on MONDAY, then 3 incremental backups (Tuesday, Wednesday, Thursday).
And if a Wednesday incremental backup is "damaged".
Can we still restore data on Thursday with using FULL backup from Monday + Incremental from Thursday?
Or we can no longer restore anything newer then full backup made on Monday because of a corrupted incremental backup from Wednesday?

4. The same as question 2. How to achieve that each new month when a new ACTIVE full backup is done that it would clean all older backups (including full and incremental).

Thank you very much for your time and help!

with best regards

[Natalia Lupacheva]

Hi Bostjan,

1. You don't have to install a backup server on this machine. You can install a repository managed by VBR and configure a Backup Copy job.
If you are going to perform this synchronization yourself via 3rd party tools, you will also need to manage your retention policy on this server.
Also, you can use our VCSP instead of the external server. You can find one in your region to make the communication simpler.

2. If you configure a periodical active full backup, the Agent will never delete the "old" active full until you reach a proper number of store points. So if you set your retention policy to 30 points, the scheme will be
- 1st of April: Active Full backup was created;
- 1st of May: new Active Full was created. The old backup data (full + incremental) is not deleted until you reach 30 restore points in May;
- 30th of May: the old backup chain is deleted (Full backup + incrementals created in April);
- 1st of June: new Active Full is created, the "old" backups are not deleted until you reach 30 restore points in June.
So you don't have to configure retention policy just to keep the full chain of incremental backups. The Agent will not delete old Active Full backup until you have the new Active Full backup + the required number of restore points.
Retention policy gives you a time gap to decide if you still need the old backup data from the previous month or not.
I believe in your case you don't have to set a retention policy to 30 points and just 3 points will be enough.

3. In this case, you will be able to restore with full backup from Monday or with full backup from Monday + incremental backup from Tuesday. As far as a backup chain is corrupted on Wednesday, you cannot restore with Wednesday's and Thursday's backup data. To restore data on Thursday you need to have Active Full + the whole backup chain.

4. In addition to previous, you can remove old data manually (if the new Active full backup is created), but it's not the best way to manage old backups. It's safer to work with retention policy as it's described in p.2.

Thanks!

[bostjanQ]

Thank you