----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/34394/#review85691 -----------------------------------------------------------
core/src/main/scala/kafka/utils/ZkUtils.scala <https://reviews.apache.org/r/34394/#comment137408> if we are going to add this it should be exposed as a configuration and written up in a KIP. We can't hard code values that folks won't understand without some clear information about why it is 5000 - Joe Stein On May 25, 2015, 3:49 a.m., Jaikiran Pai wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/34394/ > ----------------------------------------------------------- > > (Updated May 25, 2015, 3:49 a.m.) > > > Review request for kafka. > > > Bugs: KAFKA-1907 > https://issues.apache.org/jira/browse/KAFKA-1907 > > > Repository: kafka > > > Description > ------- > > KAFKA-1907 Set operation retry timeout on ZkClient. Also mark certain Kafka > threads as daemon to allow proper JVM shutdown > > > Diffs > ----- > > clients/src/main/java/org/apache/kafka/common/utils/Utils.java > f73eedb030987f018d8446bb1dcd98d19fa97331 > core/src/main/scala/kafka/network/SocketServer.scala > edf6214278935c031cf493d72d266e715d43dd06 > core/src/main/scala/kafka/server/DelayedOperation.scala > 123078d97a7bfe2121655c00f3b2c6af21c53015 > core/src/main/scala/kafka/server/KafkaServer.scala > e66710d2368334ece66f70d55f57b3f888262620 > core/src/main/scala/kafka/utils/ZkUtils.scala > 78475e3d5ec477cef00caeaa34ff2d196466be96 > > Diff: https://reviews.apache.org/r/34394/diff/ > > > Testing > ------- > > ZkClient was recently upgraded to 0.5 version, as part of KAFKA-2169. The 0.5 > version of ZkClient contains an enhancement which allows passing of operation > retry timeout https://github.com/sgroschupf/zkclient/pull/29. This now allows > us to fix the issue reported in > https://issues.apache.org/jira/browse/KAFKA-1907. > > The commit here passes the operation retry timeout while creating the > ZkClient instances. The commit was contains a change to mark certain threads > as daemon to allow a clean shutdown of the Kafka server when the zookeeper > instance has gone done first. > > I've locally tested that shutting down Kafka, after zookeeper has already > shutdown, works fine now (it tries to reconnect to zoookeeper for a maximum > of 5 seconds before cleanly shutting down). I've also checked that shutting > down Kafka first, when zookeeper is still up, works fine too. > > > Thanks, > > Jaikiran Pai > >
