Oracle ZFS Storage Appliance: AKD restart on an Active head will cause a Takeover from the Passive head

Asset ID:	1-72-2215751.1
Update Date:	2017-05-11
Keywords:

Solution Type Problem Resolution Sure

Solution 2215751.1 : Oracle ZFS Storage Appliance: AKD restart on an Active head will cause a Takeover from the Passive head

Applies to:

Oracle ZFS Storage ZS4-4 - Version All Versions to All Versions [Release All Releases]
Sun ZFS Storage 7420 - Version All Versions to All Versions [Release All Releases]
Sun ZFS Storage 7320 - Version All Versions to All Versions [Release All Releases]
Oracle ZFS Storage ZS3-4 - Version All Versions to All Versions [Release All Releases]
7000 Appliance OS (Fishworks)

Symptoms

When an AKD restart (maintenance system restart) was performed from a node that was in the active state in the clustered system, a takeover of the passive node was reported on the active node.

These are alerts that were generated on the active node.

2016-12-3 09:57:17 Show alert detailsffdd393b-ae1e-edc5-d2bd-e5f829d7d688 The appliance has taken over from its cluster peer. Minor alert

2016-12-3 09:55:27 Show alert details9c062b22-27c9-4372-8f0b-d20b1c450860 The appliance has rejoined the cluster. Minor alert

2016-12-3 09:54:16 Show alert details03ecdd48-8b95-6d8f-904f-b612bc515e31 A cluster interconnect link has been restored. Minor alert

2016-12-3 09:54:02 Show alert details67031dce-7987-e876-b601-89e1aa993629 An administrator has reset the appliance management interface or issued a diagnostic reboot. Major alert

2016-12-3 09:53:44 Show alert details70d562ba-4b51-4155-9674-830c29bcef52 A cluster interconnect link has been restored. Minor alert

2016-12-3 09:53:44 Show alert details212166c2-49e1-ee42-f1dc-f69c281f2ac2 A cluster interconnect link has been restored.

Even though a takeover was reported, nothing happened on the passive node. It did not reboot and continued to be the passive node.

Cause

A AKD restart on an active node, where the other node is passive is designed to report a takeover on the active node.

This was confusing to the customer and he thought a transfer of resources had taken place when none had.

Solution

One can monitor configuration cluster on both nodes during this sequence to see if a a true takeover has occurred.

If you believe that a node has rebooted, you can check the "Last Booted" field by clicking on the Oracle logo at the top left in the BUI

References

<BUG:25232060> - CLUSTERED NODE REBOOTED UPON RESTART OF AKD FROM THE AK CLI

Attachments

This solution has no attachment