Saw what I think is strange replication/failover behavior while testing out
a cluster with two nodes.  My cluster is running defaults and maps simple
keys to byte payloads.  I saw this sequence of events:

1) Query for key "X".  Receive the correct value "Y" as expected.

2) Shut down node #1.  Query for key "X" on node #2.  Get a timeout.  Based
on my knowledge of Riak this is expected.  Defaults are N=3 and R=2, so two
or even all three of the replicas may be on the downed node #1.  R=2
constraint can't be met.

3) Set R=1.  Query for key "X" again on node #2.  Get a response that key
"X" doesn't exist.  This behavior I don't fully understand.  My expectation
is that if node #1 contains no replicas it would timeout, and if it
contained 1 it would return the correct value "Y".

Is this expected behavior?
_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com

Reply via email to