Hmm so so I cannot replicate anything causal between the CommunicationError and the C lusterOffline: Unable to access functioning Riak node. I guess the cluster really was going offline from my benchmarking!
However, I do always a CommunicationError when Put-ing a key that has changed, when using IfNotModified = true. Just wondering if that is to be expected or maybe worth filing an enhancement request on github? Because it's not really a CommunicationError :) Alex On Wed, Oct 2, 2013 at 2:52 PM, Alex Rice <[email protected]> wrote: > Just trying to make sure I understand what is the proper usage of this > option IfNotModified in CI. > > // (first do a Get of the object) > // do a SetObject on the riakobject with my changes > var putOpts = new RiakPutOptions () { IfNotModified = true }; > // do a Put with those Options > > When the put fails, because I have forced another client to have > modified the key, I get a result back with: > > resultCode = CommunicationError > message = Riak returned an error. Code '0'. Message: modified > > So the actual condition is kind of buried in the message string which > is not ideal. Maybe there could be a first-class resultCode named > Modified? > > Also I then seem to get subsequently > ClusterOffline: Unable to access functioning Riak node > > Is it possible that CI is interpreting the CommunicationError > internally to think the node is down or something? Just curious- maybe > the unable to access functioning Riak node is because I'm running > basho_bench simultaneously? Could be a co-incidence, but it doesn't > seem like it. > Thanks, > Alex _______________________________________________ riak-users mailing list [email protected] http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
