[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.11.patch Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.10.patch, YARN-540.10.patch, YARN-540.11.patch, YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.6.patch, YARN-540.7.patch, YARN-540.7.patch, YARN-540.8.patch, YARN-540.9.patch, YARN-540.9.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.7.patch Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.6.patch, YARN-540.7.patch, YARN-540.7.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.8.patch bq. why didnt this code in the previous patch cause an exception to be thrown for a normal job? Because I added a check in RMAppRemovingTransition instead of FinalTransition bq. Can the app crash while its waiting to be unregistered. Will that generate an ATTEMPT_FAILED? Can the node crash and cause an ATTEMPT_FAILED. Since AppAttempt is already in FINISHING state if App is in REMOVING state. if app crashed, attempt will receive CONTAINER_FINISHED event and then attempt goes to FINISHED state. If the node crash, attempt should receive EXPIRE event and attempt should go to FINISHED state as well. bq. We probably need to save the previous state and return that while the app is in REMOVING state. Yes, added a function to return the previous state when App is in REMOVING state Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.6.patch, YARN-540.7.patch, YARN-540.7.patch, YARN-540.8.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.9.patch Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.6.patch, YARN-540.7.patch, YARN-540.7.patch, YARN-540.8.patch, YARN-540.9.patch, YARN-540.9.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.10.patch uploaded a new patch that removed the newly added transition for RMStateStore exception logging as that's already logged in RMStateStore Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.10.patch, YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.6.patch, YARN-540.7.patch, YARN-540.7.patch, YARN-540.8.patch, YARN-540.9.patch, YARN-540.9.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.10.patch Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.10.patch, YARN-540.10.patch, YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.6.patch, YARN-540.7.patch, YARN-540.7.patch, YARN-540.8.patch, YARN-540.9.patch, YARN-540.9.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.7.patch Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.6.patch, YARN-540.7.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.4.patch Upload a patch that changes FinishApplicationMasterResponse to contain a response-completed field and MR AM and AMRMClient are changed to retry till it becomes true. Also fixed Bikas's last comments Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.5.patch New patch fixed the test case Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.4.patch, YARN-540.5.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.3.patch upload a new patch that - removeApplicationState in RMAppAttempt.AMUnregisteredTransistion and RMApp.FinalTransition - rename RMAppEventType.ATTEMPT_FINISHING to ATTEMPT_UNREGISTERED Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.3.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Vinod Kumar Vavilapalli updated YARN-540: - Priority: Major (was: Blocker) Synced up with [~jianhe] and [~bikassaha] offline, and we all agree that the correct solution is RM restart that preserves work. If RM preserves work, then it will not blindly start new ApplicationAttempts, accepts connections from old AMs and so we are good. Reducing the priority for now to change MR to have some work-around via MAPREDUCE-5471. Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.1.patch upload a patch without tests, will add tests later on. Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Priority: Blocker Attachments: YARN-540.1.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jian He updated YARN-540: - Attachment: YARN-540.2.patch upload a new patch and add test cases to test the state machine transitions. Did single node RM restart test. will do more rigorous manual test Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Priority: Blocker Attachments: YARN-540.1.patch, YARN-540.2.patch, YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (YARN-540) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart
[ https://issues.apache.org/jira/browse/YARN-540?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Vinod Kumar Vavilapalli updated YARN-540: - Summary: Race condition causing RM to potentially relaunch already unregistered AMs on RM restart (was: RM state store not cleaned if job succeeds but RM shutdown and restart-dispatcher stopped before it can process REMOVE_APP event) Race condition causing RM to potentially relaunch already unregistered AMs on RM restart Key: YARN-540 URL: https://issues.apache.org/jira/browse/YARN-540 Project: Hadoop YARN Issue Type: Sub-task Components: resourcemanager Reporter: Jian He Assignee: Jian He Priority: Blocker Attachments: YARN-540.patch, YARN-540.patch When job succeeds and successfully call finishApplicationMaster, RM shutdown and restart-dispatcher is stopped before it can process REMOVE_APP event. The next time RM comes back, it will reload the existing state files even though the job is succeeded -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira