spindown disabled a disk, thanks unraid
Hi !
Just wanted to share my experience.
Yesterday, a power loss scheduled by my provider happened. So I wanted to turn off unraid and avoid draining batteries. So, I stopped all VMs and Containers, then, spin down all the (capable) disks. (I wanted to check UPS time left in case disks are spin down)
Then I stopped the array. I imediatly saw disk 6 disabled, amazing..
Had to rebuild the disk once power was restored, even if the disk was completly fine, and the data intact. (Unraid started the array at boot even if a disk was disabled, it was to late.. thanks unraid)
What I think happened:
Something tried to write something to disks after spin down, wich caused them to spin up. The disk took to much time to be ready, and SAS maybe "rejected the write". Unraid threated it as a write error, and tagged the disk as bad.
What drove me crazy:
- Unraid not telling me why the disk is disabled.
- Unraid trying to preserve your data by disabling a disk, then start the array even if there is something "wrong" without asking me (Array auto start is enabled)
- the fact that I can't say "No, this this is okay, everything is okay, just do a parity check if you want"
- the fact that by trying to preserve data, it rather put it in danger, because there is no solution other than rebuild, wich cause higher load on every disks
I'm okay with unraid being conservative but, come on, tell me why you flagged the disk and give me the option to overide it if I'm sure everything is okay.
And, I think I get it wrong, isn't actually the point of parity check to correct possible write errors happened ?
Btw, except those often frictions, I still really like unraid, but I also hate it sometimes.
The kind of software that is good when everything is fine but.. a different story when something go wrong.
TL;DR: Avoid spinning down disk in SAS setup if you don't want to rebuild disk. Unraid won't let you tell the disk is okay
EDIT: Everything was rebuilt in 14h and is fine now, I'm not blaming, but still wondering what happened
EDIT2 : Just wanted to add that, after using unraid for 8 years, my general experience is : everything is good until you have to stop the array and / or reboot. But when you don't, it runs rock solid for month and handle load without issues. Sadly you have to stop the array for almost any modification in settings (I get that it is to protect data, but if this then cause issues... )