Hi everyone,
I am building a FreeNAS backup server for VM data and i would like to deduplicate weekly full backups on it.
The box is an older i7-2600k CPU with 32GB of RAM, two Samsung 970Pros 512GB and 6 WD Red 4TB drives.
What ia have done already is put 4 of these WD 4TB drives in stripe with the 970Pros acting as dedupe table disks. I’ve built a dataset with 128k deduplication and put a SMB share on that, and got some 80-140 MB/s big file paste speed on that thing.
Im pretty satisfied with the results and intend to do a lot more testing with doing a RAIDZ2 on those 6 4TB disks, enabling encryption, etc.
What ia wanted to ask if anyone has suggestions how to do the thing more proprerly ?
I am unable to find quallity materials online on this idea of mine and im not sure about this new dedupe table drives feature from freenas vs. L2ARC witch stored dedupe tables before…
A lot of unknows here, so if any ZFS wizards be out there, please do leave some suggestions, i hope it will benefit not only myself due to the scarcity of quallity info.