Either shut down master or let it finish cleanup
------------------------------------------------
Key: HBASE-951
URL: https://issues.apache.org/jira/browse/HBASE-951
Project: Hadoop HBase
Issue Type: Bug
Components: master
Reporter: Jim Kellerman
Assignee: Jim Kellerman
Fix For: 0.19.0
When HMaster is called from the command line, don't try to issue a shutdown
request if it is already in the process of shutting down.
The problem is that HBaseAdmin is trying to connect to the master, and the
master has already shut down its server
threads, so it cannot respond to the HBaseAdmin.shutdown() request.
It is harmless with respect to data retention/loss, but obviously annoying from
the user's point of view.
HBaseAdmin trys to contact the master for HMaster -stop, but if the
master is already shutting down, HBaseAdmin does not know that and keeps
retrying as follows:
{code}
org.apache.hadoop.hbase.MasterNotRunningException:
content-repo.cluster.powerset.com:60000
at
org.apache.hadoop.hbase.client.HConnectionManager$TableServers.getMaster(HConnectionManager.java:221)
at org.apache.hadoop.hbase.client.HBaseAdmin.<init>(HBaseAdmin.java:67)
at org.apache.hadoop.hbase.master.HMaster.doMain(HMaster.java:799)
at org.apache.hadoop.hbase.master.HMaster.main(HMaster.java:818)
2008-10-21 18:23:35,827 INFO org.apache.hadoop.hbase.master.HMaster: Waiting
for dfs to exit safe mode...
2008-10-21 18:23:45,847 INFO org.apache.hadoop.hbase.master.HMaster: Waiting
for dfs to exit safe mode...
2008-10-21 18:23:55,858 INFO org.apache.hadoop.hbase.master.HMaster: Waiting
for dfs to exit safe mode...
{code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
