Read/write errors when running parity check

[megaz221]

I keep having issues with read/write errors on certain disks and it seems to occur during parity checks most of the time. I was getting some UDMA CRC errors on a couple of disks but it doesn't seem to have occurred in a while and not sure if it's related.

 

Disk 3 in specific seems to keep getting disabled even though SMART data looks fine other than the UDMA CRC errors. It does have about 5 years of power on time but other than that I don't see anything out of the ordinary. 

 

Disk 1 is the oldest of the array with over 5 years power on time, but has a ton of UDMA CRC errors. I'm not sure where these errors come from, I've checked cable connections and everything seems fine.

 

I have also connected the server to UPS to prevent any unclean shutdowns during a power outage, I've had a couple instances of the server shutting down uncleanly during power outages.

 

I have attached diagnostics for reference. I'm also runnnig an extended SMART test on it as we speak and it's about 50% done.

 

And yes, I'm aware the disks are very full. I've ordered two 14TB drives to add to the array but not sure if I'm just going to have to replace disk 1 and 3 (i would rather not assuming there is no actual issue).

 

 

unraid01-diagnostics-20230501-1333.zip

[megaz221]

And just for more info, I'm running unRaid as a VM in ESXi, passing through two LSI controllers that have all of the disks attached.

 

[trurl]

1 minute ago, megaz221 said:

running unRaid as a VM

Not officially supported. Have you asked in the Virtualizing Unraid subforum?

[megaz221]

5 minutes ago, trurl said:

Not officially supported. Have you asked in the Virtualizing Unraid subforum?

I have not since I kind of assumed the issue would be unrelated since i'm just passing the LSI controllers down to the VM. Was hoping someone could just look at the diagnostics and see anything out of the ordinary.

 

I can ask in that subforum if it's more appropriate.

[trurl]

Looks like a connection problem

[megaz221]

Awesome thanks for checking, I'll verify all my connections again and make sure they are good. Can you let me know where determined that from the diagnostics? I'm just curious

[trurl]

12 minutes ago, megaz221 said:

where determined that

syslog starting here:

May  1 01:11:05 unraid01 kernel: sd 3:0:1:0: [sdc] tag#453 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=DRIVER_OK cmd_age=3s

Also high UDMA CRC Error count on disk3

[JorgeB]

mpt2sas_cm0: LSISAS2008: FWVersion(20.00.02.00), ChipRevision(0x03), BiosVersion(07.39.02.00)
mpt2sas_cm1: LSISAS2308: FWVersion(20.00.07.00), ChipRevision(0x05), BiosVersion(07.39.02.00)

 

First HBA should be updated to the same firmware as the second one, that one has known issues, and possibly not coincidentally the disabled disk is connected there.

[megaz221]

Thank you so much, I'll definitely upgrade the firmware on that HBA. Extended SMART test just completed without error.

[megaz221]

I got the HBA upgraded with no issues. I'll look to add the drive back into the array and see if this fixes the issue:

 

Thanks again.

[megaz221]

Is there a way to see what disks are connected to which HBA through the unRaid console?

[JorgeB]

Tools -> System devices, then assuming no hardware changed since last diags:

0b:00.0 -> HBA1 (cm0)

13:00.0 -> HBA2 (cm1)

[megaz221]

So am i correct in assuming that the ones highlighted on the top are on HBA1 and the bottom ones are on HBA2?

[JorgeB]

Yep

[megaz221]

Thank you, really appreciate everyone's help. I'll mark this solved.