[
https://issues.apache.org/jira/browse/EAGLE-1029?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16026157#comment-16026157
]
ASF GitHub Bot commented on EAGLE-1029:
---------------------------------------
Github user qingwen220 commented on the issue:
https://github.com/apache/eagle/pull/939
This bug can be produced as follows
1. create policies with stream1 to use up all the working slot of alert
engine topology1
2. setup another alert engine topology topology2
3. create a policy with stream1 and trigger the scheduler to schedule
4. check schedule result by GET rest/metadata/schedulestates, and compare
groupSpecs & alertSpecs
> Wrong groupSpec is generated when more than one alert engine topology
> ----------------------------------------------------------------------
>
> Key: EAGLE-1029
> URL: https://issues.apache.org/jira/browse/EAGLE-1029
> Project: Eagle
> Issue Type: Bug
> Affects Versions: v0.5.0
> Reporter: Zhao, Qingwen
> Assignee: Zhao, Qingwen
> Priority: Critical
>
> According to the schedule info returned by /rest/metadata/schedulestates
> In alertSpecs, new policies are assigned to alertBolt 9, 18, 19, 1, 12 of the
> second alert engine topology. While in groupSpecs, the worker queue(alert
> bolts) allocated for each policy is not [ 9, 18, 19, 1, 12]. Actually, it
> adds all working queues, including those of other alert engine topology
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
