Parity drive in error state

[wgstarks]

Not sure what happened but my parity drive is off-line this morning and I got an alert-

Event: Unraid array errors
Subject: Warning [BRUNNHILDE] - array has errors
Description: Array has 1 disk with read errors
Importance: warning

Parity disk - ST8000AS0002-1NA17Z_Z840L3CQ (sdc) (errors 10)

Diagnostics attached.

 

brunnhilde-diagnostics-20181015-0719.zip

[JorgeB]

Parity dropped offline so there's no SMART report, check cables and/or power cycle the server so it comes back online and post new diags.

[wgstarks]

Here's the new diagnostics after reboot.

 

brunnhilde-diagnostics-20181015-0810.zip

[JorgeB]

Still offline, rebooting might not be enough:

 

58 minutes ago, johnnie.black said:

check cables and/or power cycle the server

 

[wgstarks]

Took the array off-line but the drive isn't available to re-assign. I'll check the cables this evening when I get home.

[wgstarks]

Doesn't look good for this drive. Shutdown the server and replaced the SATA cable with a new spare. When I attempted to boot I got this via IPMI-

 

 

I was able to reassign this disk as parity though. Parity sync/rebuild is currently in progress. Current smart report and diagnostics attached.

 

brunnhilde-smart-20181015-1623.zip

brunnhilde-diagnostics-20181015-1624.zip

[JonathanM]

23 minutes ago, wgstarks said:

I was able to reassign this disk as parity though. Parity sync/rebuild is currently in progress.

I wouldn't even bother. When SMART flags on boot, the drive is a goner IMHO. SMART is typically conservative on its recommendation. When it says failed, believe it.

[wgstarks]

28 minutes ago, jonathanm said:

I wouldn't even bother. When SMART flags on boot, the drive is a goner IMHO. SMART is typically conservative on its recommendation. When it says failed, believe it.

I don't see anything in the smart report though other than the UDMA CRC error count. I really don't know much about HDD's but it's my understanding that this usually indicates a bad cable.❓

Did I miss something else? Or maybe I'm wrong? Happens sometimes.😄

[JonathanM]

Dunno what to tell you. Was that SMART report taken after the image you posted? If so, that's strange, I don't remember ever seeing a drive fail SMART and subsequently pass it.

[JorgeB]

That's weird, SMART status bad means there should be a failing now SMART attribute, still disk is showing some issues, mainly:

183 Runtime_Bad_Block       -O--CK   063   063   000    -    37

If it fails again I would replace it.

[wgstarks]

@jonathanm

The SMART was taken after the boot up completed. I was surprised by it as well. Running a new one but with the parity sync running it'll take a while.

 

 @johnnie.black

Thanks. I might go ahead and check the warranty status. Might still be covered.