Immutability - retention

[frankive]

Hi.
Say we have an old VM which be accident was uploaded to an object storage with immutability.
The immutability settings is 30 days.

We now want it to be removed, so I assume we at least will have to wait for 30 days, thats fine.
But as long as the settings is in place, will each new daily backup put a flag for 30 days, so it will actually never remove the immutability flag?
Would we need to disable Immutability at the capacity repository leve for 30 days, then remove, then enable?

Thanks!

[marco.horstmann]

Hi Frank,
this restore point and is blocks are stored 30+10 days in this object storage.
The retention is set by object. Every block of a backup is stored as a separate object
in the storage. If this block is not required for any newer restore point it can be
deleted after 40 days.

Disabling immutability is no option because we used e.g. in Amazon the Compliance Mode.
Even the "admin" account unable to delete an object or remove retention for an object.

See https://helpcenter.veeam.com/docs/backu ... ml?ver=110 for more details.

Best Regards
Marco

[frankive]

Thanks!

So.. same VMs could be using the same blocks, correct?
say you have 2 almost identical servers (say 2 Citrixservers). Same VM image, OS etc.
If you would remove one of them from a immutability object storage.. doesnt the 2 vms share most of the same blocks? and 1 of the vms willl always need the blocks for its restore? Which again means I cant delete the the other VM since it uses the same blocks?

Sorry for this questions, but just trying to really understand how the immtability works

[veremin]

If you would remove one of them from a immutability object storage.

Not sure how it's possible, the immutable blocks cannot be removed by definition. Can you elaborate on your question, please?

[dalbertson]

@frankive As Vladimir stated you wouldnt be able to remove 1 of the 2 VMs until the immutability period was over anyways. But lets imagine the same situation without immutability, in the instance to where there are two VMs using the "cloned" blocks and one is "moved". The moved VM would be "rehydrated" on the new datastore and if supported on the new datastore the block cloning would start fresh there.

Does that make sense?

a copy of the block would be moved if it is still needed in the current location.

[frankive]

It makes sense, but I think I need to learn more how to visualize how the object storage works.
It seems like my understanding of "Veeam Backup Files" is not working the same in the world of object storage.

Will do some more studying in this matter and see if there is a lightbulb coming my way

If you know of some good blogs which goes a bit deeper in this matter, both Veeam related or object storage I am more than happy to have some tips..

[frankive]

say we havre 28 restore points (1 each day for a month), 12 monthly and 3 yearly backup to a SOBR with amazon s3 immutable.
Settings are copying restore points immidiately and move files older than 28 days.

Whats the best setting for the immutability setting to be sure that restore points after 28 days will get clean as usul? should it be set to 28?

[veremin]

28 will do, however, the blocks restore points consist of will not be removed immediately upon expiration, but some time after it (up to 10 days). This is additional period was introduced to reduce the solution cost by minimizing the number of requests to the object storage. More information regarding it and feature in general you can find in the UG section mentioned Marco. Thanks!

[theviking84]

Correct me if wrong but I thought there is no global dedupe still. Also sobr and object storage used per vm chains.

So that should mean that there are not blocks which are shared by two vms ever. Only that within each vm backup there will be lots of objects and blocks shared.

So vm 1 deletion should never use blocks from vm 2. I thought Gostev stated this was how it worked at one point but just want to confirm.

[Gostev]

Yes, I can confirm that.