Re: CL.ONE gives UnavailableException on ok node
Yes, if you want to keep writes available w/ RF=1 then you need to use CL.ANY. On Fri, Apr 15, 2011 at 3:48 PM, Mick Semb Wever wrote: > On Fri, 2011-04-15 at 15:43 -0500, Jonathan Ellis wrote: >> Sure sounds like you have RF=1 to me. > > Yes that's right. > > I see... so the answer here is that i should be using CL.ANY ? > (so the write goes through and hinted handoff can get it to the correct > node latter on). > > ~mck > > -- > "The fox condemns the trap, not himself." William Blake > | http://semb.wever.org | http://sesat.no > | http://tech.finn.no | Java XSS Filter > -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of DataStax, the source for professional Cassandra support http://www.datastax.com
Re: CL.ONE gives UnavailableException on ok node
On Fri, 2011-04-15 at 15:43 -0500, Jonathan Ellis wrote: > Sure sounds like you have RF=1 to me. Yes that's right. I see... so the answer here is that i should be using CL.ANY ? (so the write goes through and hinted handoff can get it to the correct node latter on). ~mck -- "The fox condemns the trap, not himself." William Blake | http://semb.wever.org | http://sesat.no | http://tech.finn.no | Java XSS Filter signature.asc Description: This is a digitally signed message part
Re: CL.ONE gives UnavailableException on ok node
Sure sounds like you have RF=1 to me. On Fri, Apr 15, 2011 at 7:45 AM, Mick Semb Wever wrote: > Just experienced something i don't understand yet. > > Running a 3 node cluster successfully for a few days now, then one of > the nodes went down (server required reboot). > After this the other two nodes kept throwing UnavailableExceptions like > > UnavailableException() > at > org.apache.cassandra.service.WriteResponseHandler.assureSufficientLiveNodes(WriteResponseHandler.java:127) > at > org.apache.cassandra.service.StorageProxy.mutate(StorageProxy.java:118) > at > no.finntech.countstats.listener.CassandraMessageListener$1.run(CassandraMessageListener.java:356) > > (this code being loosely based off the second example in > http://wiki.apache.org/cassandra/ScribeToCassandra ). > > This seems a bit weird to me when the StorageProxy.mutate(..) is being > called with ConsistencyLevel.ONE. > > I'm running 0.7.4 so i doubt it to be CASSANDRA-2069 > > ~mck > > -- > "Everything you can imagine is real." Pablo Picasso > | http://semb.wever.org | http://sesat.no > | http://tech.finn.no | Java XSS Filter > > -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of DataStax, the source for professional Cassandra support http://www.datastax.com
CL.ONE gives UnavailableException on ok node
Just experienced something i don't understand yet. Running a 3 node cluster successfully for a few days now, then one of the nodes went down (server required reboot). After this the other two nodes kept throwing UnavailableExceptions like UnavailableException() at org.apache.cassandra.service.WriteResponseHandler.assureSufficientLiveNodes(WriteResponseHandler.java:127) at org.apache.cassandra.service.StorageProxy.mutate(StorageProxy.java:118) at no.finntech.countstats.listener.CassandraMessageListener$1.run(CassandraMessageListener.java:356) (this code being loosely based off the second example in http://wiki.apache.org/cassandra/ScribeToCassandra ). This seems a bit weird to me when the StorageProxy.mutate(..) is being called with ConsistencyLevel.ONE. I'm running 0.7.4 so i doubt it to be CASSANDRA-2069 ~mck -- "Everything you can imagine is real." Pablo Picasso | http://semb.wever.org | http://sesat.no | http://tech.finn.no | Java XSS Filter signature.asc Description: This is a digitally signed message part