Wenjun Ruan

> > > > * Do you means when failover we need to take-over the previous task which has already submitted to remote cluster? > > > > > >...

> First you need to create a DSIP. Yes you need to create a DSIP, and provide detailed design, rather then directly submit a PR.

What is `OTHER_CONTINUE `, please add detail design and specific motivation.

The current naming of our strategies is confusing It's better to have a more intuitive name e.g. ``` Continue_Other_Task("One task failed will continue other running task") Continue_Other_Task_Chain("One task failed will...

Seems caused by worker failover? since this whole logic has already refactored, please test on dev branch.

> Thank you for response. On a dev branch ? What about version 3.2.2 In fact, no one maintain 3.2.2 😢

> What about 3.3.2 ? 3.3.2 is a suggested version.

> What about 3.1.9? 3.1.9 will not maintain.

Is there any error log in master? or error command in t_ds_error_command? You can get the scheduler count metrics by `ds_master_quartz_job_executed`

`ProcessScheduleTask#executeInternal` is running on master. You need to provide more information, e.g. your cluster information, is this bug can reproduce?