-
Notifications
You must be signed in to change notification settings - Fork 13.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Can not get more logs of DAGS on UI when deploy on Kubernetes #31054
Comments
I have a similar issues related to logs after upgrading to docker image 2.6.0 (I tried both helm chart 1.8.0 and 1.9.0) P.S After a task run I also see logs until |
cc: @dstandish - seems that it is a bit more widespread than just celery case. we might want to take a look closer. Can you please @chuxiangfeng @iJanki-gr describe your deployment configuration (logging configuratio, whether your logs are distrubuted and shared among multiple components (log folder persistency) and whether you use remote logging at all). |
I use official helm chart. No remote logging, each worker has its own pvc. In this instance though I have only 1 worker. The logs are present in the worker pod log directory, but webserver is not showing them for some reason. |
Do you have log persistence enabled in the chart? Only then the logs would be present in the webserver. |
Enabling log persistence is a very quick solution/workaround to the problem - not sure if final and the only one we will suggest, but it is confirmed to solve the problem |
Yes I have log persistence. Problem is still present. |
You mentioned that your webserver does not have the log folder - likely you do not have persistence properly configured because if you have log persistence enabled, the logs should appear in your webserver too.. Did you check if you have the right configuration of log persistence and the settings you have it configured with is properly set for your k8S? do you have the PVC created for logs and is it properly mounted in your wbeserver (the same PVC volume should be mounted for your worker and for your webserver if you have it properly configured) Can you double check those and see if it works for you - if you think you have log persistence enabled and you do not see the logs in the log folder in your webserver then you need to double check your configuration. Once you fix it and logs will be present in both - worker and webserver, the problem should be fixed. |
I will double check that. But I thought the webserver was connecting to the worker, and the worker was serving the logs from port 8793. |
The persistence is enabled in workers. There is no persistence configurable for webserver in the helm chart. But the log serving from worker was working fine regardless |
Yes. This is the bug that we have in 2.6.0 and will be fixed in 2.6.1 - what I proposed you was a workaround you could implement so that you do not have to wait @iJanki-gr in case you missed the context. |
I also suggest - when we release 2.6.1rc with a fix coming here: #31101 I strongly recommend that you take part in testing it and verifying if the fix works. This is the least you can do to help the community that mostly works in their free time so that you can get software for free. |
Cool, I didn't realise the issue was already identified. Sure I can test the new RC, thanks. |
I also faced that issue after upgrading to 2.6.0. Deployment of 2.6.1 did not help. Still do not see logs for the tasks in the UI. |
Just mentioning that version 2.6.1 worked for me. |
Have to apologize.... |
i‘m sorry to say after update to 2.6.1, the logs still in that state as question above. |
Reopened. Can you please see if you can provide fresh logs/more details @chuxiangfeng ? CC: @dstandish - possibly worth taking a look. |
this is all log show in ui, from 2023-06-21, 02:00:05 CST to 2023-06-21, 02:00:07 CST, pod is running, but the logs are always in loading state in ui, after pod end, after the pod runs, the ui logs are printed |
Can you please update to 2.6.2 - there were a few issues connected to logs in 2.6.1 and 2.6.2 and they seem to be related (see for "logs" in the release notes). I think they should fix your issue. Closing it provisionally - we can always reopen if upgrade does not help. |
cc: @dstandish |
sorry to say the problem still exists after the upgrade to 2.6.3. |
can you please provide more logging information and describe exactly the circumstances you have @chuxiangfeng ? That would help to focus the problem. |
in 2.7 the problem returned |
I believe (again) it does not add value to say 'it's returned' as opposed to providing more information/logs and details of the circumstances when it returned (I know it's much more time to gather all that info but helping those who might want to help to diagnose the problem and help to solve it. But maybe It's too high of an expectation for users of free software like Airflow? Not sure. Maybe they do not understand that people contribute here and help others in their free time so and it's a good idea to help them by providing more information? |
Apache Airflow version
2.6.0
What happened
After update airflow to 2.6.0 and apache-airflow-providers-cncf-kubernetes==6.1.0,Can not get more logs of DAGS on UI when deploy on Kubernetes,It was like suspended animation, but it was able to get the log in the end.
Before update,Kubernetes tasks logs will look like this in UI:
But now is likes this:
.......
(In this stage, there is no Pod startup log and no subsequent print during the Running period, so the Pod Running log cannot be recognized, In fact, Pod is running normally on k8s. f12 of Chrome has been calling the rotation training query log, but no updated log content can be obtained. At the end of the Pod it prints out Pod t-crawler-gem-ipo-request-71zd2ngd has phase Running, In theory, the print Pod t-crawler-gem-ipo-request-71zd2ngd has phase Running)
What you think should happen instead
No response
How to reproduce
Just start a KubernetesPod task.
Operating System
PRETTY_NAME="Debian GNU/Linux 11 (bullseye)" NAME="Debian GNU/Linux" VERSION_ID="11" VERSION="11 (bullseye)" VERSION_CODENAME=bullseye ID=debian HOME_URL="https://www.debian.org/" SUPPORT_URL="https://www.debian.org/support" BUG_REPORT_URL="https://bugs.debian.org/"
Versions of Apache Airflow Providers
apache-airflow==2.6.0
apache-airflow-providers-cncf-kubernetes==6.1.0
apache-airflow-providers-common-sql==1.4.0
apache-airflow-providers-dingding==3.1.0
apache-airflow-providers-elasticsearch==4.4.0
apache-airflow-providers-ftp==3.3.1
apache-airflow-providers-http==4.3.0
apache-airflow-providers-imap==3.1.1
apache-airflow-providers-mongo==3.1.1
apache-airflow-providers-mysql==5.0.0
apache-airflow-providers-postgres==5.4.0
apache-airflow-providers-redis==3.1.0
apache-airflow-providers-sqlite==3.3.2
Deployment
Official Apache Airflow Helm Chart
Deployment details
No response
Anything else
No response
Are you willing to submit PR?
Code of Conduct
The text was updated successfully, but these errors were encountered: