Improper error handling in face of version mismatch between the client and the
server
-------------------------------------------------------------------------------------
Key: HBASE-2569
URL: https://issues.apache.org/jira/browse/HBASE-2569
Project: Hadoop HBase
Issue Type: Bug
Components: client
Reporter: Benoit Sigoure
Priority: Minor
I use trunk, and every once in a while, when I upgrade my test instance of
HBase, my clients end up in a seemingly endless loop when instantiating an
{{HTable}} instance:
{code}
10/05/18 23:42:46 DEBUG zookeeper.ZooKeeperWrapper: Read ZNode
/hbase/root-region-server got 192.168.0.7:61322
10/05/18 23:42:47 DEBUG client.HConnectionManager$TableServers: Root region
location changed. Sleeping.
10/05/18 23:42:48 DEBUG client.HConnectionManager$TableServers: Wake. Retry
finding root region.
{code}
I'm pretty certain that the code throws an appropriate exception to complain
about the version mismatch but it must be getting eaten somewhere and the code
probably applies the usual retry logic instead of throwing up its hands in the
air and immediately giving up, as I think it should.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
