Deduplication - Inline and post processing

[BenBrazil]

Hi All,

In investigating some recent sis.changelog.full error messages on a couple of volumes and subsequently deduplication in general, I have a question about the differences between inline deduplication on AFF volumes and post processing deduplication on the same volume. What does the post processing deduplication do that the inline doesn’t do? I assume that the inline deduplication is limited in what it can dedupe and that the post processing deduplication is therefore more extensive? Would there ever be a reason not to run post processing deduplication and just rely on inline deduplication?

Thanks,

Ben

[paul_stejskal · Accepted solution]

Please review https://kb.netapp.com/Advice_and_Troubleshooting%2FData_Storage_Systems%2FFAS_Systems%2FHow_to_maximize_storage_efficiency_on_All_Flash_FAS_systems_ru... for the recommended setup. Inline = on-the-fly dedupe/compression/compaction, but post-process=scheduled scan of whole file system first time then changelog the second time.

To make the storage efficiency savings more efficient, please consider high change rate workloads and how you might set those up with your account team. We do not recommend disabling inline as that is a feature of AFF.

[BenBrazil]

Thanks Paul for you reply.