This message was deleted.
# general
s
This message was deleted.
b
I would expect there might be info in the overlord log. Coordinator is more for handling of segment placement, and overlord manages tasks and stuff. (Unless I just need more coffee.)
Generally, pausing is nicer than a reset, because of the offset issue. In my mind, the main difference is that a pause retains the current offset, and a reset doesn't.
I wouldn't expect it to unpause on restart, but I'd recommend testing it out to see.
a
Both restart after pause and restart after terminate continue from the last consumed offset. This is because the offsets are stored against the datasource in the metadata table.
Coordinator is also functioning as the Overlord in the cluster. So there is just one log.
b
My confusion, I guess I was thinking of a hard reset, instead of terminating and restarting.