Hi all,
I’ setting up a truenas scale server to run plex.
I’d like some advice on the best disk configuration.
The host is an i5 Intel CPU with 16GB RAM
I have 2 x 128 GB disk which i plan to use for boot in mirror
5 x 4 TB SATA disk which i was thinking configure as Raidz1 with one vDev and 2 dataset, one for the apps (Plex syncthing and FileBrowser) and 1 for the data
1 x 300 GB SATA disk which i’m not sure what to do with
the data are being backed up on a separate host.
I have space for 1 more SATA SSD disk and 2 M.2 SSD disks in case you can advice on how to best use the potential additional space
I’‘m not particularly concerned, data are backed up on a separate machine so i would need a really huge failure to loose data, and even so it won’t be an unrecoverable loss, just annoying put all together again.
And I’ considering that last failed disk I had is a 20 years old 300 GB, I feel relatively safe.
The read speed seems fairly good on RAIDZ1 so I thought that would give the best compromise between space and speed, but I’ ready to listen different configuration option.
That would need put the app on the boot dataset, I’m not against but I have some concern on Plex DB that can grow quite large.
Would add disks as cache or log add any benefit to this configuration?
Any use i can do for that spare 300 GB disk?
Because, as was explained a gazillion times in this forum and the previous one, L2ARC is not a dumb extra cache space. ZFS actually needs RAM to manage what goes and what doesn’t go in L2ARC.
The traditional recommendation for L2ARC is between 5 times and at most 10 times the size of the ARC (RAM) and at least 64 GB RAM to begin with. There was a discussion here whether this guidance could be relaxed due to evolutions in OpenZFS, which you may look for if you want; I’m personally not convinced that going “for the edge” is a good idea.