Learn how to (re)provision Tungsten Cluster datasources using the tprovision tool, from failed node recovery to full site rebuilds, while choosing backup methods that balance downtime, performance, and safety.
This session explains how tprovision runs on the target host, triggers a backup on the chosen source, streams and restores it, removes old THL, and brings the node back into the cluster, with support for mysqldump, xtrabackup, mysqlclone, and rsync.
You’ll also review when to provision from replicas vs primaries, how rsync’s two-pass seed/final sync works, key options like --source, --method, --threads, and --create-master, and how to rebuild an entire dataservice in CAA topology with a single command.