2013-03-02 10:20 AM
I have noticed on all new installs (or wipeclean defaults) that auto giveback is enabled on all nodes. In 7-mode this is always disabled and most never enable it since we want to troubleshoot any failure without autogiveback.
Does anyone run HA Pair autogiveback and why? I can think of some reasons but still haven't left it enabled.
To turn it off on all nodes in a cluster
stor fail modify -node nodename* -auto-giveback false
2013-03-03 09:10 AM
The change you have noticed is a recent one. It applies to 7-Mode and cONTAP. From the Release Notes for 8.1.2...
"Beginning with Data ONTAP 8.1.2, Data ONTAP will, by default, automatically giveback to the partner after a takeover due to the partner panicking. This giveback can be disabled or reenabled using the new -auto-giveback-after-panic parameter of the storage failover modify command .The automatic giveback reduces the amount of time the HA pair operates with a node in takeover mode."
This change was implemted after analysing support cases and observing customer experiences of takeover / giveback events.
2013-03-03 09:12 AM
Thank you... I didn't catch that in the release notes. I don't like having autogiveback though since if a hardware issue or stuck in partial giveback... but not a big concern since it does check for partner waiting for giveback.