[ https://issues.apache.org/jira/browse/BEAM-6191?focusedWorklogId=175501&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-175501 ]
ASF GitHub Bot logged work on BEAM-6191: ---------------------------------------- Author: ASF GitHub Bot Created on: 14/Dec/18 17:43 Start Date: 14/Dec/18 17:43 Worklog Time Spent: 10m Work Description: swegner commented on issue #7220: [BEAM-6191] Remove redundant error logging for Dataflow exception handling URL: https://github.com/apache/beam/pull/7220#issuecomment-447398897 > Sorry! What's the notification mechanism used here so I know to watch for it? I get GitHub notifications in my email. Note that it might be going to your personal email account, depending on how you have GitHub configured. [Here's the docs](https://help.github.com/articles/about-notifications/). > Does that OOM message actually come through? Don't think I've ever seen it, but it'd sure be handy! I don't know for sure-- I'm new to this code. I would imagine it should come through. Some reasons it wouldn't: (a) If OOM's typically manifest from some other place, or (b) if when we OOM we don't flush Dataflow logs to Stackdriver before the VM goes down. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking ------------------- Worklog Id: (was: 175501) Time Spent: 50m (was: 40m) > Redundant error messages for failures in Dataflow runner > -------------------------------------------------------- > > Key: BEAM-6191 > URL: https://issues.apache.org/jira/browse/BEAM-6191 > Project: Beam > Issue Type: New Feature > Components: runner-dataflow > Reporter: Scott Wegner > Assignee: Scott Wegner > Priority: Minor > Time Spent: 50m > Remaining Estimate: 0h > > The Dataflow runner harness has redundant error logging from a couple > different components, which creates log spam and confusion when failures do > occur. We should dedupe redundant logs. > From a typical user-code exception, we see at least 3 error logs from the > worker: > http://screen/QZxsJOVnvt6 > "Aborting operations" > "Uncaught exception occurred during work unit execution. This will be > retried." > "Failure processing work item" -- This message was sent by Atlassian JIRA (v7.6.3#76005)