You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by "Andrew MacBean (JIRA)" <ji...@apache.org> on 2014/11/06 14:23:33 UTC

[jira] [Created] (QPID-6216) Report to the BDB HA VirtualHostNode as early as possible that is's no longer the Master node

Andrew MacBean created QPID-6216:
------------------------------------

             Summary: Report to the BDB HA VirtualHostNode as early as possible that is's no longer the Master node
                 Key: QPID-6216
                 URL: https://issues.apache.org/jira/browse/QPID-6216
             Project: Qpid
          Issue Type: Bug
          Components: Java Broker
            Reporter: Andrew MacBean
            Assignee: Andrew MacBean


BDB HA VirtualHostNode currently relies on the detach event from the JE to tell it that it is no longer master.  HA acceptance testing has shown that sometimes this event might not be generated by JE for sometime after we first learn we are restarting.   Whilst this does not introduce a functional defect, it does mean the have more failed transactions and the client fails a larger number of times.  It makes the logs harder to read.

It might make Qpid more supportable if we send our own event to BDB HA VirtualHostNode when we first learn we are restarting.  We can then have the BDB VirtualHostNode close the VirtualHost early - preventing new connections and closing existing ones.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org