Crazy backup behaviour

[Joern]

Hello everybody,

I have some crazy problem using Veeam B&R v7.0.0.771 to backup some virtual machine based on ESXi 5.5.0 and vSphere 5.5.0.

I created four jobs to backup all VMs.
All jobs are configured to "incremental backup and synthetic full on friday with five restore points".

Having a look into the folders where the backup is placed, I see the following results:
1st and 4th job:

• Full on 05.03.
• four incremental backups on 7th, 8th, 9th and 10th
• 2nd Full on 12.03.
• two incremental backups on 14th and 15th

2nd and 3rd job:

• Full on 28.02.
• twelve incremental backups starting on 29th

Why are these four jobs creating different files on filesystem?

Is there a possible configuration to create this scenario?
1. Full backup
2. four incremental
3. Create synthetic full and delete 1. full to save disk storage
Is the option"transform previous..." the one I need?

Thank you
Joern

[foggy]

Joern, any chance you're using jobs chaining to execute one job after another?

[Joern]

Two jobs are running the same time and the other jobs are starting, when the first job is finished.

[foggy]

That's the reason. Your chained jobs actual start times silp to another day, so they are not started on the day the synthetic full is schedule at, hence it is not performed (regular incremental run is performed). Just one of the undesired consequences of jobs chaining, which we typically do not recommend to use.

Then, I suggest to review how retention works for forward incremental mode and decide on whether you'd want to switch to forever forward incremental or reversed incremental.

[Joern]

Thank you for your answer. Now I understand the reason of this bevavior.
I decided to create two chains is, that we get massive performance problems using a single jobs that created a backup file greater than 1 TB. Then we only get about 3MB/s instead of 82 MB/s.

Now the next question is, how to make the existing backup smaller.
Actually my disk is out of storage. Do I have to delte the whole backups before I enable the option "transform previous full backup chains into rollbacks"?
Ore d I still need to change the day of synthetic full by today to make it work?

[foggy]

Instead of chaining, you can schedule all the jobs with short intervals one after another and set max allowed number of concurrent tasks per proxy and repository servers, if required.

You can enable transform now, but it will be performed after the next synthetic full is created. To immediately reclaim some space, you can delete the oldest full backup with its corresponding increments for those jobs that have several fulls already created.

[Joern]

What is the correct way to delete the last full?
Just delete it in the filesystem?

[foggy]

Yes, you can delete it manually. However corresponding restore points will still be displayed in Veeam B&R console until removed from database according to retention.

[Joern]

Okay, thank you very mcuh.
I will try.