BCJ compression settings

[matthewingram]

We are using Nimble Hybrid arrays for our ReFS repo's mounted on Windows server 2016 servers. Following the best practice guides all of the repo's are set to Align backup file data blocks, decompress backup data blocked before storing and user per-vm backup files. The backups jobs are setup with inline dedup off, compression set to None and storage optimization set to large blocks. The backup copy jobs also have inline dedupe disabled and compression set to none. I am trying to figure out how to optimize the BCJ bandwidth. We I see any reduction in data transferred over the wan by enabling inline dedup and compression on the BCJ? if so, will the target still decompress before storing the data? What I would like is for the BCJ to compress data on the source side, send it over the wan and then decompress before storing it in the remote repository.

[HannesK]

Hello,
I believe there is some kind of typo in your question, but I try to answer it as I understand it.

The "decompress data blocks before storing" (see helpcenter) is a repository setting. So if you configure high compression on the BCJ or even WAN accelerators, then the amount of traffic will be reduced.

Repository settings and backup job / backup copy job settings are independant.

Best regards,
Hannes

[matthewingram]

That was my hope. So the repo sever on the destination side will decompress the data. As far as the traffic flow is concerned during a BCJ is a proxy used at either end or is it just repo server to repo server?

[HannesK]

the traffic flow for a backup copy job is from repo to repo. the proxy is used for backup (virtual infrastructure to repo and the other way round for VM restore)

https://helpcenter.veeam.com/docs/backu ... ml?ver=100