Data Domain inline Dedupe on both jobs?

[rntguy]

Can someone explain how inline deduplication works between backup jobs and backup and copy jobs?

We backup to windows storage first so we enabled inline dedupe. Then we do a copy job to a data domain so we disabled it.

But does the data that goes from the windows repository to the data domain get 'un-deduped' before it's sent off to the dd? or does it remain deduped further hindering the DD deduplication so I should disable inline dedupe on the backup job as well?

[foggy]

Yes, data is being "rehydrated" prior sending, so Inline dedupe setting on the original backup job doesn't affect backup copy job.

[rntguy]

thanks. so if the backup job repository has plenty of space then it would help the copy job go faster if we didn't use inline dedupe, correct? because it wouldn't have to rehydrate? but the backup job would take longer because it would no use inline dedupe?

or is the overhead of rehydration very tiny compared to the space savings of enabling inline dedupe on the backup job?

[foggy]

I called it rehydration, since similar blocks belonging to different VMs are still sent over to target. Actually, backup copy job just reads the required block according to digests and it doesn't matter whether dedupe is on or off in terms of performance. So you can leave it enabled on the backup job. Also, keep in mind that Veeam B&R uses pretty big block sizes (comparing to dedupe devices) so the space savings are typically quite small comparing to the entire VM size.

[rntguy]

thanks again. i'll leave it enabled.