On 05/30/2011 12:46 AM, sujith wrote:
Hi
Please help me. Got following error while starting a qpid cluster with
persistent store
Daemon startup failed: Cluster-ID mismatch. Stores belong to different
clusters.
To ensure data is not corrupted, qpid marks each brokers store with a UUID when
the cluster starts. If you mix stores from different clusters you'll get the
error above.
Stores are marked "dirty" while the cluster is running, so old data won't be
used after restart.
To shut the cluster down cleanly use "qpid-cluster -k". That way all brokers
will mark the stores "clean" for restart.
To start up correctly when some brokers crash, set in config:
cluster-size=N (number of members in your cluster)
On start-up the brokers will wait till all N members are present before deciding
what store to use, so they will correctly chose a clean store.
If you want to erase an old store, you need to erase the stores and cluster
identity for _all_ brokers before starting:
rm -rf /var/lib/qpidd/*
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:[email protected]
