Zookeeper DR backup needed for Flink HA mode?

Posted by David Corley on
URL: http://deprecated-apache-flink-user-mailing-list-archive.369.s1.nabble.com/Zookeeper-DR-backup-needed-for-Flink-HA-mode-tp20172.html

We're looking at DR scenarios for our Flink cluster. We already use Zookeeper for JM HA. We use a HDFS cluster that's replicated off-site, and our
high-availability.zookeeper.storageDir
property is configure to use HDFS.

However, in the event of a site-failure, is it also essential that we have a remote Zookeeper instance at the DR site thats a part of the quorum? It'd have to be non-voting due the latency involved, so possible an observer node?
https://zookeeper.apache.org/doc/r3.3.3/zookeeperObservers.html

Or is Flink smart enough that when we bring it back up in the DR location it will read from (the replicated) HDFS, register with the local ZK quorum, and continue without issue?