You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ozone.apache.org by "xBis7 (via GitHub)" <gi...@apache.org> on 2023/02/22 15:34:39 UTC

[GitHub] [ozone] xBis7 opened a new pull request, #4300: HDDS-8009. OM HA metrics should be unregistered if leader is not known

xBis7 opened a new pull request, #4300:
URL: https://github.com/apache/ozone/pull/4300

   ## What changes were proposed in this pull request?
   
   Patch #4140 added OM HA metrics, with some logic in OM to unregister if leader is unknown. However, unregistration happens in the wrong if branch, should be done if leader is null, not its id (which is required to be non-null).
   
   Related discussion: https://github.com/apache/ozone/pull/4140#discussion_r1114254142
   
   ## What is the link to the Apache JIRA
   
   https://issues.apache.org/jira/browse/HDDS-8009
   
   ## How was this patch tested?
   
   This patch was tested manually in docker clusters under 
   
   `/hadoop-ozone/dist/target/ozone-1.4.0-SNAPSHOT/compose/ozone-ha`
    and
   `/hadoop-ozone/dist/target/ozone-1.4.0-SNAPSHOT/compose/ozone`.
   
   Also, after taking another look in method `OzoneManager.updatePeerList()` under which we register `OMHAMetrics`, I can see that this method is only called after electing a leader. Under any scenario that there is no leader, this method won't even get called so we won't have to worry about RatisServer leader being null. It should be safe to remove the leader check altogether.
   
   `OzoneManager.updatePeerList()` gets used only in `OzoneManagerStateMachine.notifyConfigurationChanged(..)` if we search the logs we can verify, that it only gets called after new leader election.
   (`2023-02-22 15:05:04,865 [om2@group-D66704EFC61C-StateMachineUpdater] INFO ratis.OzoneManagerStateMachine: Received Configuration change notification from Ratis. New Peer list:
   `)
   
   ```
   2023-02-22 15:05:04,798 [grpc-default-executor-1] INFO server.RaftServer$Division: om2@group-D66704EFC61C replies to ELECTION vote request: om3<-om2#0:OK-t2. Peer's state: om2@group-D66704EFC61C:t2, leader=null, voted=om3, raftlog=Memoized:om2@group-D66704EFC61C-SegmentedRaftLog:OPENED:c0, conf=0: peers:[om1|rpc:om1:9872|admin:|client:|dataStream:|priority:0|startupRole:FOLLOWER, om3|rpc:om3:9872|admin:|client:|dataStream:|priority:0|startupRole:FOLLOWER, om2|rpc:om2:9872|admin:|client:|dataStream:|priority:0|startupRole:FOLLOWER]|listeners:[], old=null
   2023-02-22 15:05:04,837 [om2-server-thread2] INFO server.RaftServer$Division: om2@group-D66704EFC61C: change Leader from null to om3 at term 2 for appendEntries, leader elected after 42ms
   2023-02-22 15:05:04,844 [om2-server-thread1] INFO server.RaftServer$Division: om2@group-D66704EFC61C: set configuration 1: peers:[om1|rpc:om1:9872|admin:|client:|dataStream:|priority:0|startupRole:FOLLOWER, om3|rpc:om3:9872|admin:|client:|dataStream:|priority:0|startupRole:FOLLOWER, om2|rpc:om2:9872|admin:|client:|dataStream:|priority:0|startupRole:FOLLOWER]|listeners:[], old=null
   2023-02-22 15:05:04,845 [om2-server-thread1] INFO segmented.SegmentedRaftLogWorker: om2@group-D66704EFC61C-SegmentedRaftLogWorker: Rolling segment log-0_0 to index:0
   2023-02-22 15:05:04,846 [om2@group-D66704EFC61C-SegmentedRaftLogWorker] INFO segmented.SegmentedRaftLogWorker: om2@group-D66704EFC61C-SegmentedRaftLogWorker: Rolled log segment from /data/metadata/ratis/5cb24680-b9e7-3c90-a862-d66704efc61c/current/log_inprogress_0 to /data/metadata/ratis/5cb24680-b9e7-3c90-a862-d66704efc61c/current/log_0-0
   2023-02-22 15:05:04,855 [om2@group-D66704EFC61C-SegmentedRaftLogWorker] INFO segmented.SegmentedRaftLogWorker: om2@group-D66704EFC61C-SegmentedRaftLogWorker: created new log segment /data/metadata/ratis/5cb24680-b9e7-3c90-a862-d66704efc61c/current/log_inprogress_1
   2023-02-22 15:05:04,865 [om2@group-D66704EFC61C-StateMachineUpdater] INFO ratis.OzoneManagerStateMachine: Received Configuration change notification from Ratis. New Peer list:
   
   ```
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org


[GitHub] [ozone] adoroszlai commented on pull request #4300: HDDS-8009. OM HA metrics should be unregistered if leader is not known

Posted by "adoroszlai (via GitHub)" <gi...@apache.org>.
adoroszlai commented on PR #4300:
URL: https://github.com/apache/ozone/pull/4300#issuecomment-1442221989

   Thanks @xBis7 for the patch, @neils-dev for the review.
   
   > it only gets called after new leader election
   
   Yes, it seems so.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org


[GitHub] [ozone] xBis7 commented on pull request #4300: HDDS-8009. OM HA metrics should be unregistered if leader is not known

Posted by "xBis7 (via GitHub)" <gi...@apache.org>.
xBis7 commented on PR #4300:
URL: https://github.com/apache/ozone/pull/4300#issuecomment-1440269408

   @adoroszlai Can you please take a look at this PR?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org


[GitHub] [ozone] adoroszlai merged pull request #4300: HDDS-8009. OM HA metrics should be unregistered if leader is not known

Posted by "adoroszlai (via GitHub)" <gi...@apache.org>.
adoroszlai merged PR #4300:
URL: https://github.com/apache/ozone/pull/4300


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org