SMART / ZFS Problems

[cjpanici]

Hello, I have a recent proxmox installation (7.4-16). Every so often, sometimes a week, sometimes a few days...I get a slug of SMART emails about hard disks failing, also ZFS error emails. They appear to all fail at the same time, which I don't buy. When this happens it always makes the ZPOOL degraded; one drive fault due to error count. I did replace some of the initial hard drives, but I am starting to think this is software related. I have one single disk that has the OS installed on it, then a ZFS pool (RAID 10) that holds the guest machine disk images, ISOs, Templates, basically everything else.

Here is my hardware specs:



I tried putting my HBA in a different slot thinking that it may be the PCI connection to the hard drive adapter card. I've attached the syslog during a time where this problem reared its ugly head. It happened 10/22 at about 3am, and I am seeing weird stuff about hard disks going into "sleep" mode? Is that a thing? Maybe something just needs to be tweaked in my bios? I did verify that the HBA is in IT mode. I've never experienced an OS crash, however, the OS disk drive is in a different array (separate cable) of the HBA. Anyway, any help is very greatly appreciated.

Thanks,
Chris

 

[LnxBil]

What disks are this? We had regular problems with SATA disks in huge zpools that randomly fail, this never occured with SAS disks.

 

[cjpanici]

These are WD Gold disks. They are connected in hot swap bays. Was that your question or were you asking about logical volumes?

 

[Dunuin]

Maybe the whole HBA is failing (old age/overheating)? Had this with a SAS2008.

 

[cjpanici]

It is a new HBA, I did end up replacing it though. I also moved all of the drives from the Front of the case to the back thinking it may be the front drive controller board that is the culprit, or the cable going from the HBA to the front drive controller board. We'll see. Thanks for the help.