[
https://issues.apache.org/jira/browse/KNOX-2807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17608644#comment-17608644
]
Sandor Molnar commented on KNOX-2807:
-------------------------------------
Additional information: it's not only the SSL_ENABLED flag is stored on the
role level instead of the service level, but the name we use is the "related
name" we read out from the CM API response (e.g.
[https://CM_SERVER:7183/api/v30/clusters/Cluster%201/services/HIVE_ON_TEZ-1/config?view=full).|https://cm_server:7183/api/v30/clusters/Cluster%201/services/HIVE_ON_TEZ-1/config?view=full).]
However, in the PollingConfigurationAnalyzer Knox checks the "name" fields
only (current vs. previous) and does not consider related names.
> Restart of HIVE_ON_TEZ causes a Knox topology redeploy
> ------------------------------------------------------
>
> Key: KNOX-2807
> URL: https://issues.apache.org/jira/browse/KNOX-2807
> Project: Apache Knox
> Issue Type: Bug
> Reporter: Sandor Molnar
> Assignee: Sandor Molnar
> Priority: Critical
> Fix For: 2.0.0
>
>
> A restart of HIVE_ON_TEZ triggers a topology redeploy: gateway.log contains
> the following line, even though there was no configuration change.
> {code:java}
> Role property hive.server2.use.SSL value has changed from true to null {code}
> Possibly caused by
> [HiveServiceModelGenerator|https://github.infra.cloudera.com/CDH/knox/blob/7a69a5128fa8052e24bae3bb012f48fad1eaf154/gateway-discovery-cm/src/main/java/org/apache/knox/gateway/topology/discovery/cm/model/hive/HiveServiceModelGenerator.java#L77]
> pushing the service level *hive.server2.use.SSL* setting to the role level,
> while
> [PollingConfigurationAnalyzer|https://github.infra.cloudera.com/CDH/knox/blob/7a69a5128fa8052e24bae3bb012f48fad1eaf154/gateway-discovery-cm/src/main/java/org/apache/knox/gateway/topology/discovery/cm/monitor/PollingConfigurationAnalyzer.java#L515]
> not doing so.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
