When the target node is down and the replication factor is 1. Any batch
mutate/read with consistency level QUORUM will fail to succeed.
----------------------------------------------------------------------------------------------------------------------------------------
Key: CASSANDRA-2064
URL: https://issues.apache.org/jira/browse/CASSANDRA-2064
Project: Cassandra
Issue Type: Improvement
Components: Core
Affects Versions: 0.7.0
Reporter: Mikael Sitruk
When the target node is down and the replication factor is 1. Any batch mutate
or read (with consistency level QUORUM) will fail to succeed.
This case occurs because the quorum cannot be verified.
The storage proxy will write the data on its own log, but cannot assure the
quorum (since the quorum definition requires the replica to respond) - the
responseHandler.assureSufficientLiveNodes(); will throw UnavailableException.
While it is OK to send an exception i think that the exception should contain
an appropriate message - like consistency level: <Quorum>, replication factor:
<replication factor> cannot be verified for key... (adding the natural and the
hinted node should be nice too at least in the log). Perhaps a separate
exception should be thrown (sub type of UnavailableException)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
