Answer
Use ./mtctl dump --namespace [namespace] — it produces a dump folder with logs for every pod, which is the way to identify which specific pod is stuck and holding up the rest (pods in this kind of deployment typically start up one at a time, and if one errors, subsequent pods won't start either). Common root causes surfaced this way include insufficient memory and database connectivity problems — one confirmed case turned out to be conflicting IP addresses between the database host and local network addressing, resolved by correcting the IP configuration and rerunning the installer.