Seeing as no one has brought it up... thinking I'm the first to fix this incredibly silly issue with Unraid.
Cloudflare(1) saw latency issues years ago with dm-crypt and implemented no_write_workqueue and no_read_workqueue which was upstreamed in Kernel 5.9 (with external patches for lower versions, of course). Using xfs-encryption on my 8 core system results in a nice system load of atleast 40, if not 120 (yes, 120, on a 8! thread system) with crazy WA times - making the system completely unusable.
As Unraid doesn't use /etc/crypttab, the only way to do this is to set the persistent flags on the volume. To apply this configuration to your disks is very simple.
cryptsetup --perf-no_read_workqueue --perf-no_write_workqueue --allow-discards --persistent refresh md8
To check your volumes, spin through `ls /dev/mapper`.
Yes, right now these disks are all rotational and there's a massive reduction in system load. No, I haven't played with read/writes or other ciphers to see if there's even more performance to get out of this. My system is usable again with a modicum of array writing so I'm moving on for now. The system load is even worse without encryption from it what seems like, ironically.
Cheers.
https://blog.cloudflare.com/speeding-up-linux-disk-encryption/
Recommended Comments
There are no comments to display.
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.