This issue has happened 3 times over the last month. It required having to cleanup the bin logs on both um1 and um2 to get the system back in service with both UMs.

I don't know if this is a BUG, but thought it need to be bugged for development info.

Issue -

mysqld on UM2 crashes after a restartSystem is perform. restartSystem is perform just as a maintenance requirement, not due to any issue. at restart time, UM2 mysqld crashes during the process of replication of the bin-log files to catch up.
But for some reason, the bin-log files are not insync with UM1 binary logs and the InnoDB database contained in idbdata1

To resolved as a work-around:

1. disabled UM2
2. startsystem with UM1 working
3. On UM1, run 'get master status' and check the bin-log shown there is the current one. So the bin-log repliation is update-to-date
run 'purge binary logs before 'today-date'
4. on UM2, move or delete the bin and replay log files
scp idbdata1 from UM1 to UM2
5. enable UM2

At this time, the system is back insync and mysqld isnt crashing.