Open ahululu opened 1 month ago
Maybe the helm cluster 3.2.1 was upgraded to 3.2.2 once before, but there is a problem, So, I use helm rollback , but the database will not roll. Now my workaround is to manually delete the operator column and then re-execute the file manually:: https://github.com/apache/dolphinscheduler/blob/a5061eb3518fd9ea5db4a85892da61781baac04c/dolphinscheduler-dao/src/main/resources/sql/upgrade/3.2.2_schema/mysql/dolphinscheduler_ddl.sql
Maybe the helm cluster 3.2.1 was upgraded to 3.2.2 once before, but there is a problem, So, I use helm rollback , but the database will not roll. Now my workaround is to manually delete the operator column and then re-execute the file manually:: https://github.com/apache/dolphinscheduler/blob/a5061eb3518fd9ea5db4a85892da61781baac04c/dolphinscheduler-dao/src/main/resources/sql/upgrade/3.2.2_schema/mysql/dolphinscheduler_ddl.sql
But I'm not sure that's the right thing to do
Maybe the helm cluster 3.2.1 was upgraded to 3.2.2 once before, but there is a problem, So, I use helm rollback , but the database will not roll. Now my workaround is to manually delete the operator column and then re-execute the file manually:: https://github.com/apache/dolphinscheduler/blob/a5061eb3518fd9ea5db4a85892da61781baac04c/dolphinscheduler-dao/src/main/resources/sql/upgrade/3.2.2_schema/mysql/dolphinscheduler_ddl.sql
Yes. You can try this manully.
After a few days of use after the upgrade, a new issue was discovered. api pod does not recognize alert pod with the following error message:
`
Caused by: java.net.UnknownHostException: dolphinscheduler-alert-xxxxxx-xxx
`
The temporary workaround for now is to add alert pod parsing to /etc/hosts .
However, this operation is problematic, and if api pod restarts, the alert function will not be available again. Does anyone know how to fix it?
cc @Gallardot
After a few days of use after the upgrade, a new issue was discovered. api pod does not recognize alert pod with the following error message:
`
Caused by: java.net.UnknownHostException: dolphinscheduler-alert-xxxxxx-xxx
`
The temporary workaround for now is to add alert pod parsing to /etc/hosts .
However, this operation is problematic, and if api pod restarts, the alert function will not be available again. Does anyone know how to fix it?
I need more error logs
After a few days of use after the upgrade, a new issue was discovered. api pod does not recognize alert pod with the following error message:
`
Caused by: java.net.UnknownHostException: dolphinscheduler-alert-xxxxxx-xxx`
The temporary workaround for now is to add alert pod parsing to /etc/hosts . However, this operation is problematic, and if api pod restarts, the alert function will not be available again. Does anyone know how to fix it?I need more error logs
It's the same problem as this issue:
Search before asking
What happened
ENV: EKS1.29
when upgrade to 3.2.2 from 3.2.1 (replace helm image), get error from pod dolphinscheduler-db-init-job-xxxx
What you expected to happen
I don't think db init scripts are well compatible with this kind of image upgrade
How to reproduce
just deploy 3.2.1 by helm chart, then upgrade to 3.2.2
Anything else
No response
Version
3.2.x
Are you willing to submit PR?
Code of Conduct