I have this issue with the xtraDB Cluster set on a docker cluster.
Got a 3 nodes cluster set up on 3 different docker engine machines.
The cluster is working fine. and whilst the DB is pretty small now it’s gonna grow.
I started running fail/recovery tests knowing that there’s an automated system in PCX that logs the last GID for IST recovery.
but, every time I restart the node it gets a -1 GID and runs a full SST.
As far as I know the gchache.size is big enough to hold the needed transactions for the IST.
Anyone got an idea why the IST doesn’t kick into action?