jj_uk Posted November 22, 2022 Share Posted November 22, 2022 I have a data disk that is showing smart errors. The disk is about 12 months old, and annoyingly a 3-cycle preclear didn't show any errors, but that's life. The drive was ripped from a USB backup box so it has no warranty. Bye bye cash. My server is running low on storage, so I've bought two 20TB drives (usb backup drives, currently connected via usb and preclearing 3x cycles). All data disks and the parity disk are currently 8TB. To insert the 20T disks, I'll also need to replace the cache drive as it must be >= the largest data disk. Would I be OK swapping out the parity drive with a 20T, rebuilding the array then swapping the failing drive with the old parity drive? Or should I do a the parity swap procedure instead ? (https://wiki.unraid.net/The_parity_swap_procedure) Quote Link to comment
JonathanM Posted November 23, 2022 Share Posted November 23, 2022 Attach the diagnostics zip to your next post in this thread if you want more accurate advice, but given what you have posted, you need the parity swap procedure. The size of the cache drive is irrelevant, but the parity drive(s) must be as large or larger than any single data drive in the array. You are planning on attaching the new drives directly to SATA, not USB, right? Quote Link to comment
jj_uk Posted November 23, 2022 Author Share Posted November 23, 2022 The mention of the cache drive was a mistake, I was supposed to be referring only to the parity and data drives. Yes- Once they've done a few weeks of preclearing on the binhex-preclear docker, i'll pop them out of the usb cases and attach then via sata. Quote Link to comment
jj_uk Posted December 8, 2022 Author Share Posted December 8, 2022 (edited) I followed the procedure 'exactly' to perform the parity swap. The old parity drive was 8T, the new parity drive is 20T. I have the following drives (image below) : The parity swap process completed successfully. The last part of the instructions states that running a parity check is optional. So I chose to run it. There are literally millions of sync errors being corrected by the parity check! - What's going on ?! The parity check position is past 8T (my largest data drive). Is this simply the parity drive being reset to a known state and nothing to worry about? The new parity drive was precleared before use. Should I be worried?? tower1-diagnostics-20221208-1820.zip Edited December 8, 2022 by jj_uk Attached diagnostics Quote Link to comment
JorgeB Posted December 8, 2022 Share Posted December 8, 2022 1 hour ago, jj_uk said: There are literally millions of sync errors being corrected by the parity check! - What's going on ?! If the errors started after the 8TB mark it's an old bug, though it's a pesky one because it only happens sometimes, nothing to worry about, just let it correct. Quote Link to comment
jj_uk Posted December 8, 2022 Author Share Posted December 8, 2022 OK, I'll stop worrying now. Seems to be reading and writing only the new parity drive. Was this fixed in a newer Unraid version (I'm using 6.10.2) ? I'm planning on replacing a data drive with a 20T once this parity check finishes. Quote Link to comment
JorgeB Posted December 9, 2022 Share Posted December 9, 2022 Still not fixed because it's difficult to consistently reproduce. Quote Link to comment
jj_uk Posted December 12, 2022 Author Share Posted December 12, 2022 The process of changing the parity drive to a 20T drive, and replacing a faulty disk has now completed. I've now replaced another small data disk with a 20T disk. As I now have loads of free space on the server, I'd like to remove a disk. The process to remove a drive is to reset the config, choosing to preserve assignments, changing the disk to be removed to 'unassigned' then starting the array (in maintenance mode, for speed). As a disk has been removed, what happens here? Does Unraid try recreate the removed disk's data onto the remaining disks? In my case, there will definitely be enough space for this to happen, but what happens if there's not enough space? What happens to the share settings if they're set to use, or exclude that disk? This isn't covered in the FAQ. https://wiki.unraid.net/Manual/Storage_Management#Removing_data_disk.28s.29 Quote Link to comment
itimpi Posted December 12, 2022 Share Posted December 12, 2022 2 minutes ago, jj_uk said: As a disk has been removed, what happens here? Does Unraid try recreate the removed disk's data onto the remaining disks? Unraid will not move data. It is up to you to manually put the data onto the other drives or perhaps use the unBalsnce plugin. Quote Link to comment
Frank1940 Posted December 13, 2022 Share Posted December 13, 2022 On 11/22/2022 at 4:33 PM, jj_uk said: The drive was ripped from a USB backup box so it has no warranty. Bye bye cash. I would suggest that you go to the manufacturer's website and check the drive's serial number for the drive's warranty status. (I seem to recall hearing that many of those drives do have a warranty and that an RMA will be issued.) Quote Link to comment
jj_uk Posted December 14, 2022 Author Share Posted December 14, 2022 On 12/12/2022 at 10:25 PM, itimpi said: Unraid will not move data. It is up to you to manually put the data onto the other drives or perhaps use the unBalsnce plugin. So the data is just lost? ok, i'm moving the data off the drive i'd like to remove using the Dynamix File Manager plugin. I'll take a look at unBalance. Quote Link to comment
Frank1940 Posted December 14, 2022 Share Posted December 14, 2022 IF that disk (that is to be removed) is still in the parity-protected array, make sure that you are copying disk-to-disk! A copy operation between a disk (in the array) and a User Share can result in data loss. Stated again, disk-to-Share (or Share-to-disk) can easily result in a data loss... Quote Link to comment
jj_uk Posted December 14, 2022 Author Share Posted December 14, 2022 40 minutes ago, Frank1940 said: IF that disk (that is to be removed) is still in the parity-protected array, make sure that you are copying disk-to-disk! A copy operation between a disk (in the array) and a User Share can result in data loss. Stated again, disk-to-Share (or Share-to-disk) can easily result in a data loss... It's copying from /mnt/disk2/MediaServer to /mnt/disk1/MediaServer So that's disk to disk, right? /mnt/user/MediaServer <-- is that a user share? Quote Link to comment
Frank1940 Posted December 14, 2022 Share Posted December 14, 2022 11 minutes ago, jj_uk said: It's copying from /mnt/disk2/MediaServer to /mnt/disk1/MediaServer Yes-- You are fine! 11 minutes ago, jj_uk said: /mnt/user/MediaServer <-- is that a user share? Yes 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.