MSA 2040 - Killed partner controller. (reason: Unreliable disk read path detected by p

[mr_vaughn]

MSA 2040 SAS with D2700 DAS. All Firmware's are up to date. 3 x ESXi hosts in place.

Controller A keeps on failing. and another disk is also reporting issues with a Recovery Error & Meduim error

here are the logs.

-----------------------------------------------------------------------------------------------

Killed partner controller. (reason: Unreliable disk read path detected by partner controller)

EVENT ID:#B34373

EVENT CODE:84

EVENT SEVERITY:Warning

EVENT TIME:2022-11-14 07:45:52



Midplane Serial Number:confidential info erased

System Name:AMC-MSA2040

System Location:LA

System Contact:suppport@innovativecomp.com

IP Address:172.16.4.208

Bundle A Version:GL225P002-02

MC A Version:GLM225R009-01

SC A Version:GLS225R06-01

Bundle B Version:GL225P002-02

MC B Version:GLM225R009-01

SC B Version:GLS225R06-01


Additional Information:
  The controller that logged this event forced the partner controller to fail over.

-----------------------------------------------------------------------------------------------

[ArunKKR]

Hi, The unreliable disk read path crash could be due to scrub errors.  Please check events for any scrub errors.  Execute below commands to disable scrub temporarily:

set advanced-settings background-disk-scrub off

set advanced-settings background-scrub off

Check whether the controller crashes stops.

Replace the disks with medium errors. 

If the second disk is reporting numerous recovered errors, replace it as well.

Enable scrub and check whether scrub gets completed without any errors. Check whether controller crashes has stopped. If controller still crashes you would need to log an HPE support case for in depth analysis. 

[ArunKKR]

Hi, You might need to take a back up, delete Pool A/vdisks owned by controller A and recreate them if issue persists. I would suggest to log a support case with HPE if issue persists after trying the initial suggestions.

[mr_vaughn]

seems tuning off the scrubbing and then replacing any disks that reported media errors have stabilized the failing A contoller. Just turned on the scrubbing and then will see if the contrller fails again..

FYI the MSA 2040 SAn is no longer available for support renewals..

and getting a supported end to end SAS on MSA2060 with Gen9 servers is not supported. So it looks like we are going Fiber channel on MSA2060