Urbanpixels Posted March 12, 2020 Share Posted March 12, 2020 Hi Guys, I have 2x cache drives (m.2) in my Unraid Pool. I think one of the drives is failing as im seeing loads of warnings. If the devices are in Raid1 i should be able to just power down the machine and replace the damaged drive yes? i will not lose any data? I expect i will just need to re-balance afterwards? Or am i totally wrong with that? Many thanks. Quote Link to comment
JorgeB Posted March 12, 2020 Share Posted March 12, 2020 To make a direct replacement you'd need both old and new devices connected, which might not be an option with m.2 devices, you can still remove one device, convert pool to single, then add another device and reconvert to raid1, but before starting make sure your pool is redundant because of this bug, if you want post diags to confirm. Quote Link to comment
Urbanpixels Posted March 12, 2020 Author Share Posted March 12, 2020 (edited) 4 minutes ago, johnnie.black said: To make a direct replacement you'd need both old and new devices connected, which might not be an option with m.2 devices, you can still remove one device, convert pool to single, then add another device and reconvert to raid1, but before starting make sure your pool is redundant because of this bug, if you want post diags to confirm. Thanks, I assumed Raid1 was a complete mirror so im surprised to learn you can't just replace one drive. Mine looks fine here i believe - Device size: 1.86TiB Device allocated: 62.06GiB Device unallocated: 1.80TiB Device missing: 0.00B Used: 53.68GiB Free (estimated): 926.11GiB (min: 926.11GiB) Data ratio: 2.00 Metadata ratio: 2.00 Global reserve: 20.12MiB (used: 0.00B) Data Metadata System Id Path RAID1 RAID1 RAID1 Unallocated -- -------------- -------- --------- -------- ----------- 1 /dev/nvme0n1p1 30.00GiB 1.00GiB 32.00MiB 922.84GiB 2 /dev/nvme1n1p1 30.00GiB 1.00GiB 32.00MiB 922.84GiB -- -------------- -------- --------- -------- ----------- Total 30.00GiB 1.00GiB 32.00MiB 1.80TiB Used 26.72GiB 120.50MiB 16.00KiB Edited March 12, 2020 by Urbanpixels Quote Link to comment
JorgeB Posted March 12, 2020 Share Posted March 12, 2020 1 minute ago, Urbanpixels said: you can't just replace one drive. Like mentioned you can, but to do a direct replacement old one needs to be connected, if you start without one of them the pool will be converted to single profile, you can then add another device. Quote Link to comment
JorgeB Posted March 12, 2020 Share Posted March 12, 2020 It's possible you could do a direct replacement without the old drive, but IIRC this could sometimes not work correctly, and I can't test at the moment. Quote Link to comment
Urbanpixels Posted March 12, 2020 Author Share Posted March 12, 2020 Perfect. Apologies i misread your initial post about the cache pool converting to single automatically. Quote Link to comment
JorgeB Posted March 12, 2020 Share Posted March 12, 2020 I did a quick test and direct replacement without the old drive should work, but you need to disable array auto-start, or array will start automatically with the single cache device (and convert to single profile) before you can add the new one, still this is not well tested so proceed with care. 1 Quote Link to comment
JonathanM Posted March 12, 2020 Share Posted March 12, 2020 3 hours ago, johnnie.black said: I did a quick test and direct replacement without the old drive should work, but you need to disable array auto-start, or array will start automatically with the single cache device (and convert to single profile) before you can add the new one, still this is not well tested so proceed with care. I think I or someone else made a feature request years ago where autostart could be optionally (by default) disabled if any array storage device changes were detected. Does that sound familiar? If not, I think it should be implemented. Quote Link to comment
JorgeB Posted March 12, 2020 Share Posted March 12, 2020 2 minutes ago, jonathanm said: I think I or someone else made a feature request years ago where autostart could be optionally (by default) disabled if any array storage device changes were detected. Does that sound familiar? Yes, I also remember asking for that on one of the release threads, IIRC it would need to be changed by LT since it's an emhttp function, it doesn't autostart if an array device is missing, but it does for a missing pool device, if there's a feature request for that might be a good idea to bump it. 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.