One of them removed itself

[luca63]

Good morning everyone.I have a big problem. In fact I keep having problems that I don't understand what they depend on.
I have a pool called hdd and it consists of 4 disks.
One of them removed itself as you can see from the picture

My questions are : how come it removed itself ? having two disks in this condition can I change only one ( the one with more errors ) ?

 

[leesteken]

My questions are : how come it removed itself ? having two disks in this condition can I change only one ( the one with more errors ) ?

Disk (electronics) died or connection got loose. Or SATA controller port/connection died. Another two drives are also returning read and checksums errors, which are typically drive problems (corrupt data) or connection problems or SATA controller problems (corrupting data in flight) or system memory issues. Possibly drives are breaking each other due to vibrations.
Do some hardware troubleshooting (which is not Proxmox specific), like memtest86 to check the system memory and SMART self-test of drives.

 

[Dunuin]

And draid is only recommended for LOTS of disks. Not if you only want to use 5 of them:
https://pve.proxmox.com/wiki/ZFS_on_Linux#_zfs_draid

dRAID is intended for more than 10-15 disks in a dRAID. A RAIDZ setup should be better for a lower amount of disks in most use cases.

And those disks are cheapest consumer HDDs not designed for any raid or 24/7 operation and they are 13-14 years old...I wouldn`t expect them to be reliable...

I would check smartctl for bad sectors and check that there are proper backups available.

 

[gfngfn256]

While I agree with Dunuin that these disks are basically Ewaste (I personally wouldn't even use them in a personal PC - forget Draid + parity disk!), however assuming almost all of the OPs disks are malfunctioning together, I'm going to bet that he has some HW issue. OP has not disclosed how they are physically connected/controlled. But my guess there's some real wanky type connection being used - causing errors + disconnection.

 

[Dunuin]

Even if they are connected to some proper HBA, if that HBA is as old as the disks, I wouldn't wonder if that HBA is failing because of old age. Had one of my LSI SAS2008 HBAs fail one year ago and all disks attached to it where showing errors until it some weeks later completely failed and wasn't even shown any longer by lspci.

 

[luca63]

tank for every answer