[
https://issues.apache.org/jira/browse/AMBARI-14572?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Siddharth Wagle updated AMBARI-14572:
-------------------------------------
Attachment: AMBARI-14572.patch
> Add zookeeper tick time setting to AMS for embedded mode
> --------------------------------------------------------
>
> Key: AMBARI-14572
> URL: https://issues.apache.org/jira/browse/AMBARI-14572
> Project: Ambari
> Issue Type: Task
> Components: ambari-metrics
> Affects Versions: 2.0.0
> Reporter: Siddharth Wagle
> Assignee: Siddharth Wagle
> Priority: Critical
> Fix For: 2.2.1
>
> Attachments: AMBARI-14572.patch
>
>
> Found out that we are not setting property
> "hbase.zookeeper.property.tickTime" for embedded mode.
> The AMS system initialization was stuck on Zookeeper being unresponsive for a
> few minutes right at the start of the cluster.
> The possible cause is still unknown, only explanation is a long gc pause .
> However, Zookeeper session timedout on start. with the following message:
> {code}
> 2016-01-05 10:24:10,000 INFO [SessionTracker] server.ZooKeeperServer:
> Expiring session 0x152114fb0060001, timeout of 40000ms exceeded
> {code}
> We set the session timeout to 120 seconds however the server session expires
> in 20 * _default_tick_time_ (2 seconds)
> We can definitely increase this 3 fold by setting a property which should
> alleviate this situation.
> *Note*: After auto-start brings up the daemon a second time, this does not
> reoccur.
> {code}
> 2016-01-05 10:23:24,906 INFO [main] zookeeper.ZooKeeper: Initiating client
> connection, connectString=localhost:61181 sessionTimeout=120000
> watcher=master:484800x0, quorum=localhost:61181, baseZNode=/ams-hbase-unsecure
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
