Benoit Tellier created JAMES-3726:
-------------------------------------

             Summary: Allow relaxing CassandraMailRepository LWT usage
                 Key: JAMES-3726
                 URL: https://issues.apache.org/jira/browse/JAMES-3726
             Project: James Server
          Issue Type: Improvement
          Components: cassandra, MailStore & MailRepository
    Affects Versions: 3.7.0
            Reporter: Benoit Tellier
             Fix For: 3.8.0


At a customer place I discovered CassandraMailRepository extensive relies on 
lightweight transactions, which turns out not playing well when doing massive 
error recovery operations on mail repositories:


{code:java}
2022-03-10T11:09:40,137 [ERROR] o.a.j.m.i.JamesMailSpooler - Failed 
Mail1645005443013-a8c3d070-afd7-4821-990d-7c309ff38970 processing 1 consecutive 
times. Mail is requeued with increased failure count.
com.datastax.driver.core.exceptions.WriteTimeoutException: Cassandra timeout 
during CAS write query at consistency LOCAL_SERIAL (2 replica were required but 
only 1 acknowledged the write)
        at com.datastax.driver.core.Responses$Error$1.decode(Responses.java:87)
        at com.datastax.driver.core.Responses$Error$1.decode(Responses.java:65)
        at 
com.datastax.driver.core.Message$ProtocolDecoder.decode(Message.java:297)
        at 
com.datastax.driver.core.Message$ProtocolDecoder.decode(Message.java:268)
        at 
com.datastax.shaded.netty.handler.codec.MessageToMessageDecoder.channelRead(MessageToMessageDecoder.java:88)
        ... 25 common frames omitted
Wrapped by: com.datastax.driver.core.exceptions.WriteTimeoutException: 
Cassandra timeout during CAS write query at consistency LOCAL_SERIAL (2 replica 
were required but only 1 acknowledged the write)
        at 
com.datastax.driver.core.exceptions.WriteTimeoutException.copy(WriteTimeoutException.java:113)
        at 
com.datastax.driver.core.Responses$Error.asException(Responses.java:167)
        Suppressed: java.lang.Exception: #block terminated with an error
                at 
reactor.core.publisher.BlockingSingleSubscriber.blockingGet(BlockingSingleSubscriber.java:99)
                at reactor.core.publisher.Mono.block(Mono.java:1703)
                at 
org.apache.james.mailrepository.cassandra.CassandraMailRepository.store(CassandraMailRepository.java:74)
                at 
org.apache.james.transport.mailets.ToRepository.service(ToRepository.java:79)
                at 
org.apache.james.mailetcontainer.impl.camel.CamelProcessor.process(CamelProcessor.java:75)
                at 
org.apache.james.mailetcontainer.impl.camel.CamelMailetProcessor$MailetContainerRouteBuilder.handleMailet(CamelMailetProcessor.java:176)
                at 
org.apache.james.mailetcontainer.impl.camel.CamelMailetProcessor$MailetContainerRouteBuilder.lambda$configure$0(CamelMailetProcessor.java:153)
{code}

In my opinion these LWT are far from relevant and could be replaced by an 
idempotent behaviour. We should expose a configuration option allowing to do 
just that.

Proposal: cassandra.properties

{code:java}
mailrepository.strong.consistency=false
{code}




--
This message was sent by Atlassian Jira
(v8.20.1#820001)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to