8 drives with read errors at the same time?

[newerNan]

Hi,

 

Never had read errors before, dont know what to do. Any help would be greatly appreciated.

Unraid is reporting that 8 drives have suddenly encounted read errors. I find that somewhat hard to believe.

 

Some remarks:

• Unassigned drives is suddenly showing all 8 of these drives with new sd[x] id's with some sort of failed array icon, while they appear to still be in the array.
• My genuine Unassigned drives arent even showing up any more (maybe a limit to how many the dashboard can display)
• Only 2 random 8TB drives didnt have any errors, every other array drive (inc. 2xParity) did.
• 6TB drives all have 1036 Errors, and 8TB+ have 1133, suggesting that these failed reads were across all drives at certain ways through the data. But i think it was fine this morning, so happened today at somepoint. I havent done a partiy check in over a month, so why would all drives be read at the same time?
• Some drives are connected via an HBA, some direct to motherboard controller. Parity is definitely on motherboard, 6TBs are definitely on HBA, so it doesnt appear to be a controller issue, because affected drives are spread between both.
• im running 6.10.3
• Between 2pm-3pm today, I did attempt to start a years old VM accidently, that failed miserably erroring on all sorts of issues. It wasnt something i expected to work, it would have been attempting to pass through GPUs that didnt exist, and used a drive image (or potentially a physical disk) for the OS that were no longer present. I feel this might have been around the time the read errors happened, but i dont understand why launching a VM would affect the array's data drives.

 

Please help. I want to ensure my data is safe, my offsite backup hasn't been working for about 2 months, have been meaning to sort it out.

Tempted to restart the server if it just resolves a one-off issue, and sorts out the unassigned drives plugin, but dont want to lose any diagnostics data.

 

Any assistance appreciated, thanks,

 

 

harmzserv-diagnostics-20221031-1515.zip

[newerNan]

Sorry, i forgot to mention, that everything appears to still be working fine. Dockers are fine, remote array access over SMB is fine, but not tested in anger.

[JorgeB]

Almost all disks disconnected and reconnected at the same time, including both parity drives which are using a different controller, so most likely a power problem, even if you do anything else for now you should at least reboot.

[newerNan]

Thanks for your reply @JorgeB

I definitely wouldn't have expected a power issue, its a highend Corsair PSU, running off a UPS, but ok, stranger things have happened.

 

New issue, stopping the array or trying to shutdown appears to do absolutely nothing. The log is 100% full at the moment, perhaps related to it not responding to these actions, but i also can't figure out how to clear that down without a reboot.

Without any new log entries to view, i have no idea why its not stopping the array or shutting down.

 

Any advice to get a clean shutdown?

 

Thanks,

[newerNan]

Nevermind, physical power button has appeared to triger a safe shutdown.

Restarted now, everything looks good and normal. Will kick off a parity check to be safe.

 

Thanks for the assistance.