[
https://issues.apache.org/jira/browse/HDDS-1836?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16890636#comment-16890636
]
Shashikant Banerjee edited comment on HDDS-1836 at 7/23/19 3:02 AM:
--------------------------------------------------------------------
[~arp], in test environment, with a limited set of dns, let's say one dn
instance could not connect to its peers. In such cases as well, the leader
election will take much longer time to fail/timeout and hence dns will not be
freed up to be part of different/new pipeline and hence we can run into a
situation where pipeline creation is not possible/delayed in such setups.
Also, we have a client request timeout of by 3s by default. If the leader
election min timeout is considerably larger than the request timeout, we can
see multiple request timeout with a single leader election window.
was (Author: shashikant):
[~arp], in test environment, with a limited set of dns, let's say one dn
instance could not connect to its peers. In such cases as well, the leader
election will take much longer time to fail/timeout and hence dns will not be
freed up to be part of different/new pipeline and hence we can run into a
situation where pipeline creation is not possible/delayed in such setups.
Also, we have a client request timeout of by 3s by default. If the leader
election min timeout is considerably larger than the request timeout, we can
see multiple request timeout with a single leader election window.
> Change the default value of ratis leader election min timeout to a lower value
> ------------------------------------------------------------------------------
>
> Key: HDDS-1836
> URL: https://issues.apache.org/jira/browse/HDDS-1836
> Project: Hadoop Distributed Data Store
> Issue Type: Bug
> Components: Ozone Datanode
> Affects Versions: 0.5.0
> Reporter: Shashikant Banerjee
> Assignee: Shashikant Banerjee
> Priority: Major
> Labels: pull-request-available
> Fix For: 0.5.0
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> The default value of min leader election timeout currently is 5s(done with
> HDDS-1718) by default which is leading to leader election taking much longer
> time to timeout in case of network failures and leading to delayed creation
> of pipelines in the system. The idea is to change the default value to a
> lower value of "2s" for now.
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
