Difference between Snapshot only backup and Storage Snapshot retention

[kanand72]

I have a small confusion.
As far as I can make out from the documentation, there are 2 ways to keep storage snapshots --- One by making them through a Storage Snapshot only backup job and the other by configuring the storage as secondary repository.

I created 2 jobs -- one in each method.
When I saw the snapshots that were created by both jobs, they were different.
In the Snapshot only backup job, I selected only one VM. When the job completed, I could see the snapshot of the entire LUN with all the VMs. But against only one VM, it was written protected by.

In the second job where I specified the primary storage as the secondary repository, in the snapshot that was created, I could only see one VM.

Why the difference or is my understanding wrong ?

Regards
Anand

[Andreas Neufert]

A bit hard to follow.
But in general it works the following way. We process snapshots and use the information collected in the job to temporary list the VMs in our UI for it. But we do not trust this information as between we have collected the information from vCenter and the Storage Snapshot initiated there could have someone implemented a change like VM delete operation. Therefore we mount the snapshot later during regular rescan to our Proxy and check what VMs are in the Snapshot and we would correct the entry in our UI. So over time you will likely see the same VMs listed on all snapshots. Some marked as consistent some as crash consistent.

When you go to the History - System you can see the background scans.

[kanand72]

So let me repeat my understanding ----
When I create a Snapshot only backup job, which contains only one VM, the storage snapshot that gets created actually contains all the VMs that are present in the LUN and not just that one VM.
I had created the job with Application aware processing enabled...So only the VM which I listed was application aware processed. The other VMs were not processed with app aware .
What is the difference between consistent and crash consistent ??

[Andreas Neufert]

Storage Snapshots always contain all the data (from all VMs) present on the VM.
If we process only one VM within a job and create a snapshot we might only know about the single VM (it depends on Job process type) and would list only temporarly that this snapshot has only one VM in your example. Next rescan will analyse what VMs are together on the snapshot and then update the snapshot UI entry and list all VMs.

This is independant of guest processing or no guest processing (consistent/unconsisten snapshot).
If you process only some VMs with consistency, then only those VMs will be marked as consistent.

[kanand72]

Got it.
Thanx.

Did a quick test . I have a 1.0 TB production VM.
Took a snapshot backup. Then did an Instant restore and a Migration.
Instant recovery took around 3-4 minutes and Migration of datastore took around 8 minutes...Thats pretty fast for a 1 TB VM restore.

[Andreas Neufert]

Nice to hear. THanks for the feedback.