Data Domain Backup Copy compression settings

[adr67]

Hi,

We have backup jobs on local proxy storage and backup copy of these jobs on Data Domain over ddboost.
I've set on Backup Copy job on compression level "Dedupe-friendly" as recommanded for this kind of deduplication appliance.

But what level of compression I have to set for the "primary" backup job ?
"Dedupe-friendly" --> We lose some disk space on local proxy storage because we won't have on optimal compression level
or
"Optimal" --> We optimise disk space on local proxy storage but what will the deduplication impact on Data Domain ?

On the Data Domain repository configuration I have the "Decompress backup data blocks before storing" activated. But I don't know if it mean that backup copy will decompress Backup job files before storing it on Data Domain backup copy repository ?

Thank you for your help guys !

[Shestakov]

Hi and welcome to the community!
I would recommend using "optimal" compression for the local repository and no compression for the secondary repository(dedup appliance) since you have the mentioned "Decompress backup data blocks before storing" option.
P.S. you can find a number of interesting related topics in the forums, such like this one

[adr67]

Thank you guy !
So during backup copy processing, vbk and vib files (from backup jobs) will be decompressed on the fly on the Data Domain ?
I could even use the maximum compression level on the local repository (I won't do it), there won't have any impact on deduplication ratio on Data Domain (exept of cpu load on proxy during decompression process) ?

[veremin]

So during backup copy processing, vbk and vib files (from backup jobs) will be decompressed on the fly on the Data Domain ?

Target datamover deployed on repository itself or on a gateway server (as is in your case, I believe) will do this, indeed.

I could even use the maximum compression level on the local repository (I won't do it), there won't have any impact on deduplication ratio on Data Domain (exept of cpu load on proxy during decompression process) ?

Correct.

Thanks.

[Shestakov]

I could even use the maximum compression level on the local repository (I won't do it), there won't have any impact on deduplication ratio on Data Domain (exept of cpu load on proxy during decompression process) ?

Vladimir`s answer is correct, but keep in mind that even "High" compression level provides just additional 10% compression ratio over the Optimal level at the cost of about 10x higher CPU usage.

[foggy]

And just to add for better understanding, backup job and backup copy job compression settings are independent, so you can set Optimal on the original job and High for the copy job (to reduce data transfer over a slow link, for example) and, if repository is set to decompress data, it will land on the storage in raw format.

[adr67]

Thanks !
Everything is clear now

[adr67]

Hi,

I've another question regarding the "enable inline data deduplication" option.
I've disable this option on backup copy jobs (pointing on data domain repository).
But should I enable this option on primary backup job ?
A bout Data Domain deduplication ratio, is there any impact to enable this option on the primary backup job (local repository) ?
Thanks

[foggy]

Yes, it is recommended to have this option enabled for the primary backup job. No impact on the DD dedupe rate.

[adr67]

Thanks!

[jmmarton]

For this scenario, wouldn't it be easier to go with the defaults? Optimal compression for the backup job, auto for the BCJ which means it will follow whatever the backup job is using, and then use decompress data blocks setting on the DD repo? Seems like the other tweaks will make things more complex than they need to be here.

Joe