Comprehensive data protection for all workloads
Post Reply
filippoAdb
Enthusiast
Posts: 34
Liked: never
Joined: Apr 08, 2021 12:59 pm
Contact:

Broken backup chain with datastore copy

Post by filippoAdb »

Case #08140997
I have a backup job which uses a local HP storeonce repository. Then a backup copy job (datastore copy) creates the same backup files on a remote HP storeonce.
The retention for datastore copy doesn't consider full or incremental backups, it deletes files older than the fixed number of days, so the oldest restore point on remote repository may be an incremental backup. I can select it for restore but the operation clearly fails saying "Full backup file for this object is not present in the backup repository". From the user point of view Veeam is telling that there is a valid restore point, but if you select it you find out that the backup chain is broken, so it is completely useless and gives you a false security.
david.domask
Product Manager
Posts: 3752
Liked: 913 times
Joined: Jun 28, 2016 12:12 pm
Contact:

Re: Broken backup chain with datastore copy

Post by david.domask »

Hi filippoAdb,

Yes, I see your point, but I was under the impression that such points would be marked as invalid in the Properties section, are you seeing differently? Basically my expectation is we should flag such points to avoid that "surprise" exactly as you found, so if you're seeing differently, let me know and will check internally on expected behavior.

The Catalyst Copy retention is a bit different, yes, and I find that this section explains how to best plan for the current implementation.
David Domask | Product Management: Principal Analyst
filippoAdb
Enthusiast
Posts: 34
Liked: never
Joined: Apr 08, 2021 12:59 pm
Contact:

Re: Broken backup chain with datastore copy

Post by filippoAdb »

So it seems to me that very few people in the world are using datastore copy.
After 2 weeks of analysis the conclusion is that it is the expected behavior to have some orphaned .vib in the secondary repository which apparently seem normal.
The problem happens when the retention policy on the backup job is less than Keep secondary copies for additional x days on datastore copy target.
When an entire backup chain is deleted on source repository enforcing the retention policy of the backup job, on target repository only those files older than x days are deleted, no matter they are .vib or .vbk. As a consequence on secondary repository you see some restore points you cannot use and in the worst case you can restore only from ( x days - (backup chain length -1)).
Post Reply

Who is online

Users browsing this forum: Amazon [Bot] and 957 guests