[jira] [Commented] (AIRFLOW-1667) Remote log handlers don't upload logs on task finish
[ https://issues.apache.org/jira/browse/AIRFLOW-1667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16364775#comment-16364775 ] Josh Bacon commented on AIRFLOW-1667: - Thanks for the explanation on internals [~ashb]. I made some incorrect assumptions, my issue was unrelated and is now resolved. Logs shipping correctly. > Remote log handlers don't upload logs on task finish > > > Key: AIRFLOW-1667 > URL: https://issues.apache.org/jira/browse/AIRFLOW-1667 > Project: Apache Airflow > Issue Type: Bug > Components: logging >Affects Versions: 1.9.0, 1.10.0 >Reporter: Arthur Vigil >Priority: Major > > AIRFLOW-1385 revised logging for configurability, but the provided remote log > handlers (S3TaskHandler and GCSTaskHandler) only upload on close (flush is > left at the default implementation provided by `logging.FileHandler`). A > handler will be closed on process exit by `logging.shutdown()`, but depending > on the Executor used worker processes may not regularly shutdown, and can > very likely persist between tasks. This means during normal execution log > files are never uploaded. > Need to find a way to flush remote log handlers in a timely manner, but > without hitting the target resources unnecessarily. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (AIRFLOW-1667) Remote log handlers don't upload logs on task finish
[
https://issues.apache.org/jira/browse/AIRFLOW-1667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16358757#comment-16358757
]
Ash Berlin-Taylor commented on AIRFLOW-1667:
The process that writes to the log files is a sub-process of the celery worker
itself – that just invokes {{airflow run --local}} - and that means the flush
should happen as soon the task instance finishes running.
I do not see this behaivour on Py3/1.9.0 - our tasks appear in S3 when the task
instance is finished. Are you saying you have to stop the {{airflow worker}}
process for the logs to appear in S3?
> Remote log handlers don't upload logs on task finish
>
>
> Key: AIRFLOW-1667
> URL: https://issues.apache.org/jira/browse/AIRFLOW-1667
> Project: Apache Airflow
> Issue Type: Bug
> Components: logging
>Affects Versions: 1.9.0, 1.10.0
>Reporter: Arthur Vigil
>Priority: Major
>
> AIRFLOW-1385 revised logging for configurability, but the provided remote log
> handlers (S3TaskHandler and GCSTaskHandler) only upload on close (flush is
> left at the default implementation provided by `logging.FileHandler`). A
> handler will be closed on process exit by `logging.shutdown()`, but depending
> on the Executor used worker processes may not regularly shutdown, and can
> very likely persist between tasks. This means during normal execution log
> files are never uploaded.
> Need to find a way to flush remote log handlers in a timely manner, but
> without hitting the target resources unnecessarily.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (AIRFLOW-1667) Remote log handlers don't upload logs on task finish
[ https://issues.apache.org/jira/browse/AIRFLOW-1667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16358749#comment-16358749 ] Josh Bacon commented on AIRFLOW-1667: - +1 We are using CeleryExecutors and notice that our logs never ship unless we shut down our workers. Flush probably needs to happen on some interval or task event handler. > Remote log handlers don't upload logs on task finish > > > Key: AIRFLOW-1667 > URL: https://issues.apache.org/jira/browse/AIRFLOW-1667 > Project: Apache Airflow > Issue Type: Bug > Components: logging >Affects Versions: 1.9.0, 1.10.0 >Reporter: Arthur Vigil >Priority: Major > > AIRFLOW-1385 revised logging for configurability, but the provided remote log > handlers (S3TaskHandler and GCSTaskHandler) only upload on close (flush is > left at the default implementation provided by `logging.FileHandler`). A > handler will be closed on process exit by `logging.shutdown()`, but depending > on the Executor used worker processes may not regularly shutdown, and can > very likely persist between tasks. This means during normal execution log > files are never uploaded. > Need to find a way to flush remote log handlers in a timely manner, but > without hitting the target resources unnecessarily. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (AIRFLOW-1667) Remote log handlers don't upload logs on task finish
[ https://issues.apache.org/jira/browse/AIRFLOW-1667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16195759#comment-16195759 ] Allison Wang commented on AIRFLOW-1667: --- Great I didn't realize the closed flag is removed in other PR. > Remote log handlers don't upload logs on task finish > > > Key: AIRFLOW-1667 > URL: https://issues.apache.org/jira/browse/AIRFLOW-1667 > Project: Apache Airflow > Issue Type: Bug > Components: logging >Affects Versions: 1.9.0, 1.10.0 >Reporter: Arthur Vigil > > AIRFLOW-1385 revised logging for configurability, but the provided remote log > handlers (S3TaskHandler and GCSTaskHandler) only upload on close (flush is > left at the default implementation provided by `logging.FileHandler`). A > handler will be closed on process exit by `logging.shutdown()`, but depending > on the Executor used worker processes may not regularly shutdown, and can > very likely persist between tasks. This means during normal execution log > files are never uploaded. > Need to find a way to flush remote log handlers in a timely manner, but > without hitting the target resources unnecessarily. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (AIRFLOW-1667) Remote log handlers don't upload logs on task finish
[ https://issues.apache.org/jira/browse/AIRFLOW-1667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16195609#comment-16195609 ] Arthur Vigil commented on AIRFLOW-1667: --- AIRFLOW-1676 and AIRFLOW-1682 actually fix an underlying problem with the remote log handlers that was causing logs to _never_ be uploaded, so the problem isn't quite as bad as I thought. It would still be nice if we could replicate the behavior of the old logging system in getting task logs pushed in a timely manner as right now the timing is unpredictable. > Remote log handlers don't upload logs on task finish > > > Key: AIRFLOW-1667 > URL: https://issues.apache.org/jira/browse/AIRFLOW-1667 > Project: Apache Airflow > Issue Type: Bug > Components: logging >Affects Versions: 1.9.0, 1.10.0 >Reporter: Arthur Vigil > > AIRFLOW-1385 revised logging for configurability, but the provided remote log > handlers (S3TaskHandler and GCSTaskHandler) only upload on close (flush is > left at the default implementation provided by `logging.FileHandler`). A > handler will be closed on process exit by `logging.shutdown()`, but depending > on the Executor used worker processes may not regularly shutdown, and can > very likely persist between tasks. This means during normal execution log > files are never uploaded. > Need to find a way to flush remote log handlers in a timely manner, but > without hitting the target resources unnecessarily. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
