[HBASE-412] unable to report to master

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 0.2.0
• Component/s: regionserver
• Assignee: Unassigned
• Reporter: Billy Pearson
• Created: 06/Feb/08 01:58
• Updated: 22/Oct/08 18:06
• Resolved: 23/Apr/08 23:00

Description

I have been getting these messages on my region server when they are under load

FATAL org.apache.hadoop.hbase.HRegionServer: unable to report to master for 30927 milliseconds - aborting server

Seams something is blocking the lease check in or something like that
I know this is not a network issue as this has happened on a region server on the same node as the master.
also I have updated hbase.regionserver.handler.count to 100 to make sure that was not the issue and the doc say that that is used for the master so its not lack of open handles to the master.
kind of a pain becuase it crashes the region server causing the hlogs to be used to restore

This is the messages right after the above and then the server shuts down if we can not find whats goiong on can we at lease make the server shutdown safely flushing memcache and closeing all logs so we do not have to do a restore on the next server that gets assigned the regions.

INFO org.apache.hadoop.ipc.Server: Stopping server on 60020

Issue Links

• is related to: HBASE-616 " We slept XXXXXX ms, ten times longer than scheduled: 3000" happens frequently.