[ 
https://issues.apache.org/jira/browse/RATIS-707?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16951540#comment-16951540
 ] 

Siddharth Wagle commented on RATIS-707:
---------------------------------------

Yea [~szetszwo] I see your point, revert is ok. How about if the minTimeout = 0 
only the first time server checks before asking for vote, irrespective of 
weather leader gets elected or not?

> Test failures caused by minTimeout set to zero
> ----------------------------------------------
>
>                 Key: RATIS-707
>                 URL: https://issues.apache.org/jira/browse/RATIS-707
>             Project: Ratis
>          Issue Type: Bug
>          Components: server
>    Affects Versions: 0.5.0
>            Reporter: Siddharth Wagle
>            Assignee: Siddharth Wagle
>            Priority: Major
>             Fix For: 0.5.0
>
>         Attachments: RATIS-707.01.patch
>
>
> TestRaftAsyncWithGrpc#testBasicAppendEntriesAsync and other tests fail if the 
> initial minTimeout is 0 then the server can trigger a leader election much 
> more frequently because the heartbeat interval is still at minTimeoutMs/2
> {code}
> 2019-10-11 00:45:47,813 INFO  impl.FollowerState 
> (FollowerState.java:run(108)) - s0@group-C51B0F2AC202-FollowerState: change 
> to CANDIDATE, lastRpcTime:21ms, electionTimeout:17ms
> 2019-10-11 00:45:47,870 INFO  impl.FollowerState 
> (FollowerState.java:run(108)) - s0@group-C51B0F2AC202-FollowerState: change 
> to CANDIDATE, lastRpcTime:35ms, electionTimeout:31ms
> 2019-10-11 00:45:47,933 INFO  impl.FollowerState 
> (FollowerState.java:run(108)) - s0@group-C51B0F2AC202-FollowerState: change 
> to CANDIDATE, lastRpcTime:51ms, electionTimeout:51ms
> 2019-10-11 00:45:47,969 INFO  impl.FollowerState 
> (FollowerState.java:run(108)) - s0@group-C51B0F2AC202-FollowerState: change 
> to CANDIDATE, lastRpcTime:22ms, electionTimeout:21ms
> {code}



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to