In Roaming SC deployment, if split-brain occurs that separates all nodes apart, in which each partition has one SC, we have all SCs becoming active. At rejoin, all SCs detect themself as duplicated active to one of other SCs, they should all reboot, ideally.
However, sometimes the last active SC is not detected as duplicated because all the other SCs already reboot. The last SC does not find any others as active duplicated to itself. As of this result, since the last SC is not healthy throughout the split time, it's causing many errors for other nodes to rejoin again after reboot.
Diff:
commit 68fde36133a5fd47b667c6971c967a7cf8629b03
Author: Minh Chau minh.chau@dektech.com.au
Date: Wed May 26 21:05:12 2021 +1000
commit ca0cb78a03a2eb3cfa3519b4c5d9af0905f325a5
Author: Minh Chau minh.chau@dektech.com.au
Date: Wed May 26 21:05:12 2021 +1000
commit bbe47278c2499bc738bf0c2dc8cc4ebbbb9a026d
Author: Minh Chau minh.chau@dektech.com.au
Date: Tue Jul 13 18:00:41 2021 +1000