Recently Kibana has been complaining about routing allocation settings preventing index upgrades, even when they're in a state which was OK for previous Kibana versions.
The elasticsearch cluster has cluster routing allocation incorrectly set for migrations to continue
My usual upgrade procedure on clusters without automation is to use the Cerebro padlock to set routing allocation to new primaries, upgrade & restart the node, and then unlock the padlock again when the node is back in the cluster.
As doing this sets routing allocation to all, Kibana gets upset and refuses to upgrade its index, because it's now expecting routing allocation settings to be absent. The fix for this is to manually set shard allocation to null to allow it to continue.
Tl;dr could we please have Cerebro set routing allocation to null instead of all when unlocking via the padlock icon?
Hi there,
Recently Kibana has been complaining about routing allocation settings preventing index upgrades, even when they're in a state which was OK for previous Kibana versions.
The elasticsearch cluster has cluster routing allocation incorrectly set for migrations to continue
My usual upgrade procedure on clusters without automation is to use the Cerebro padlock to set routing allocation to
new primaries
, upgrade & restart the node, and then unlock the padlock again when the node is back in the cluster.As doing this sets routing allocation to
all
, Kibana gets upset and refuses to upgrade its index, because it's now expecting routing allocation settings to be absent. The fix for this is to manually set shard allocation tonull
to allow it to continue.Tl;dr could we please have Cerebro set routing allocation to
null
instead ofall
when unlocking via the padlock icon?