In mysql-galera cluster , sometimes while performing certain activities which affects the connection ,
the value of the variable wsrep_local_recv% stuck in all nodes of the cluster.
To rectify this all nodes needs to evicted and again the whole cluster is reformed which requires a heavy downtime if there is huge data.This downtime is not affordable in production environments.
Is there any way to rectify this issue without evicting the nodes and reforming the cluster?
It is really a painful issue.
Please provide a solution as this is the critical issue.
I'm afraid I'm not familiar with such issue (unless it's something else which I can't figure out based on your explanation). If you are a ClusterControl user, please do open a ticket with us and we will try to help you.Comment actions
Please sign in to leave a comment.