je82 Posted January 6 Share Posted January 6 One of the biggest downsides to unraid for me is the fact that if there were to be any parity sync errors its very difficult to know what files may have been affected by the error. I don't know if there is a simple solution to this, i guess you could use the hash plugin to create a unique hash for every file and then have it compare the hash on all files and if it has changed then there's the problem.... but this requires a massive amount of cpu power and time if you have many files/big array. I'm not by any means smart, my hopes is that the brilliant people behind unraid can build a feature that can detect what files have gone bad from corruption in a better more efficient way? Quote Link to comment
itimpi Posted January 6 Share Posted January 6 If this really worries you then it is probably easiest to instead use btrfs (or possibly ZFS) as the format for the array disks. This had built-in check-summing to detect errors in files, but has the downside of being thought to be more fragile when hardware errors occur. Quote Link to comment
Bastian Posted January 6 Share Posted January 6 (edited) 4 hours ago, itimpi said: use btrfs (or possibly ZFS) as the format for the array disks Did some research and stumbled upon this post on SE. Tl;dr: For zfs, redundant configuration (mirror, raidz, raidz2, raidz3) can correct themself at read, single disks have to set copies to 2 or higher to correct errors. The part about striped I didn't understand 😕 Edited January 6 by Bastian Quote Link to comment
itimpi Posted January 6 Share Posted January 6 2 minutes ago, Bastian said: Did some research and stumbled upon this post on SE. Tl;dr: For zfs, redundant configuration (mirror, raidz, raidz2, raidz3) can correct themself at read, single disks have to set copies to 2 or higher to correct errors. The part about striped I didn't understand 😕 Not sure what you are saying? You asked about detecting corruption - not about correcting it when found. Quote Link to comment
Bastian Posted January 6 Share Posted January 6 (edited) 12 minutes ago, itimpi said: Not sure what you are saying? You asked about detecting corruption - not about correcting it when found. I am not op, just a by-passer adding his 5 cents ^^ Reading your post I was curious how zfs would behave in that scenario and posted my findings. True, I might have missed the original questions a bit 😅 Edited January 6 by Bastian Quote Link to comment
Recommended Posts
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.