Finally, syncLimit, specifies the max synchronization time between a follower and the leader. The second, initLimit, defines how long follower ZooKeeper servers can try to synchronize with the leader ZooKeeper for the first connection. The first parameter defines the value of 1 tick in milliseconds. tickTime, initLimit and syncLimit - the main time unit in ZooKeeper is called tick.In both nf (local) and global_nf (configuration store) you will find: In fact, they're very similar and by themselves, don't show what's are the differences between the stored data. Let's see know what are the differences between the configuration files of these ZooKeeper instances. So no worries, global-zookeeper is the previous name for configuration-store, probably kept for retrocompatibility. In fact, it's normal because DEFAULT_GLOBAL_ZK_CONF disappeared in 2018 in the quite self-explanatory commit "Hide deprecated options and few more adjustments". How are they set? In bin/pulsar we can find the default configuration only for the latter entry:ĭEFAULT_CONFIGURATION_STORE_CONF=$PULSAR_HOME/conf/global_nf global-zookeeper uses PULSAR_GLOBAL_ZK_CONF whereas configuration-store works on PULSAR_CONFIGURATION_STORE_CONF. global-zookeeper vs configuration-storeīoth commands use the same class but the configuration is different. Well, 2 classes and 3 ZooKeeper-like commads? It sounds strange, so let's see what's the difference between "global-zookeeper" and "configuration-store" commands. PULSAR_LOG_FILE=$ -Dreadonlymode.enabled=true"Įxec $JAVA $OPTS $ASPECTJ_AGENT =$PULSAR_LOG_FILE .ConfigurationStoreStarter $PULSAR_CONFIGURATION_STORE_CONF that you can see 2 classes responsible for ZooKeeper management, respectively ZooKeeperStarter and ConfigurationStoreStarter. To discover what's going on, I analyzed bin/pulsar file and looked for "zookeeper" inside. The second one works across all clusters and is called configuration store. The first one is called local and it operates at the cluster level. In the documentation section about Apache ZooKeeper in Apache Pulsar you can find the information about 2 different ZooKeepers. In the next one, I will check why we configure 2 localities in 3 different ways whereas finally, I will explore the configuration options of both ZooKeepers. In the first one, I will focus on the ZooKeeper locality regarding Pulsar brokers. This blog post is composed of 3 sections.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |