Pillar Axiom: Pilot Replacement May Result in Pilot OS Not Being Updated Properly (Ax600)

Asset ID:	1-72-1391492.1
Update Date:	2012-09-04
Keywords:

Solution Type Problem Resolution Sure

Solution 1391492.1 : Pillar Axiom: Pilot Replacement May Result in Pilot OS Not Being Updated Properly (Ax600)

Applies to:

Pillar Axiom 600 Storage System - Version Not Applicable to Not Applicable [Release N/A]
Information in this document applies to any platform.
This issue only affects Ax600 systems without the fix for defect 58308.

Symptoms

Pilot CU replacements are manufactured and stocked at the lowest software revision supported by the specific pilot hardware. As a Pilot CU is replaced, the surviving Pilot CU will update the replacement to match the currently installed software and firmware versions to those of the currently installed release. If the versions do not match, the surviving Pilot CU will install its entire image and then reboot the new Pilot CU to bring it up on that matched image.

If the replacement fails, the Support --> Software Configuration --> Software Modules screen may indicate multiple installed versions for Pilot OS and/or Pilot Software. This may cause problems if there is a Pilot CU failover or may prevent future software upgrades from completing successfully.

Cause

On the Ax600, a logic error in the checks used to make sure the software components are correctly updated before rebooting the new Pilot CU may fail. As a result, the new Pilot may not have the correct Pilot OS or Pilot Software version installed. This may cause problems if there is a Pilot CU failover or may prevent future software upgrades from completing successfully.

Solution

Please contact the Support Center to determine if your installed release contains the fix before replacing pilot CUs. The Support Center can work with an Administrator Level 1 account to enable SSH console access to your pilot and install a workaround to allow pilot replacement.

If the pilot CU has already been replaced, please contact the Customer Support center for recovery.

Both the workaround and the recovery are non-disruptive to data access

IMPORTANT!

When replacing a Pilot CU, always make sure all cables, except AC power, are securely attached to the new Pilot CU before applying power. The Pilot software uses communication on the serial and ethernet network to determine pilot role and identity. If power is applied before these cables are securely attached, a split brain situation may occur which could result in a system down condition.

Common Questions

Question: What are the consequences of this issue?

Answer: This depends on the specific software version installed on your Axiom. If the replacement Pilot CU is not properly updated, it may not be able to manage the Axiom if the other Pilot CU fails. This may result in a system down condition.

Question: If the Pilot CU has already been replaced, is it possible to recover?

Answer: Yes, but please do not attempt any recovery without Support Center assistance. Check the Support --> Software Configuration --> Software Modules GUI screen and note whether there are multiple installed versions of Pilot OS and/or Pilot Software.

Question: Is this a disruptive process?

Answer: The workaround prior to Pilot CU replacement is not disruptive. The recovery process is also non-disruptive. If the Pilot CU has been replaced, and there is a Pilot CU failover for any reason, that could cause a disruption of service. Therefore, if the Pilot CU has been replaced and the Pilot OS was not updated, please contact the Support Center for recovery as soon as possible.

Question: What is involved in Pilot SSH Console access?

Answer: For security, an Administrator Level 1 account must modify the "pillar" account to allow SSH Access. Then a special SSH Access license (and Support Tool) must be installed, and the "pillar" account must then issue a supporttool request to enable access.

Attachments

This solution has no attachment