Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Later
-
None
-
None
-
None
-
None
Description
When working on HBASE-7205, I logged the following comment:
The reason for coprocessor loading to tolerate ClassNotFoundException is that the user table region where jar file and classname mismatch can reside on any of the region servers in the cluster. If we don't tolerate the mismatch, it is possible that all region servers in the cluster abort.
I think we should allow region server UI to display critical errors such as those corresponding to ClassNotFoundException when loading user coprocessors. This would help user quickly identify issues without forcing them to scan region server log(s).
Region server can also send such critical errors to master so that master can aggregate the errors and give user a central place to check critical errors.