Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
Zookeeper >> mail # user >> Is quorum lost when leader goes down?


Copy link to this message
-
Re: Is quorum lost when leader goes down?
On Mon, Oct 1, 2012 at 3:34 PM, Jordan Zimmerman
<[EMAIL PROTECTED]> wrote:
> If the leader crashes I would assume that quorum is lost for a short time until a new leader can be elected. Is that true?

yes.

> What is the client's view of this? Does the client see an connection disruption when the leader crashes?

All servers stop serving client connections during re-election. The
client tcp connections are all closed. Once a new leader is elected
and the followers are all in sync the servers start serving client
connections again.

Also see the "reliability" section here:
http://zookeeper.apache.org/doc/r3.4.4/zookeeperOver.html#Reliability

Patrick
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB