Hello all, I've opened an issue here: https://jira.mariadb.org/browse/MDEV-33720 I just hope someone in the community can help me figure out what's wrong with this. In a 3-node galera cluster, one of the node is a replication slave from remote master, while the 2 others are simply synced with the first. When I restart mariadb service on those 2 nodes, it simply doesn't restart at all and there's no logs shown anywhere. It sits there for hours. /usr/sbin/mariadbd --user=mysql --wsrep_recover --disable-log-error If I restart the first node (the slave of master cluster), there is no issue it restarts as expected. Whether I stop the slave replication doesn't make any difference. It used to work fine, so I'm not sure what changed since updating to 10.6.16, then .17 or if it's related. We have 2 clusters and the other cluster, same configuration, is not suffering from this, but it's the master. Is there any way to investigate this further?