* Optimize master log, add workflow instance id and task instance id in log
* Use MDC to set the workflow info in log4j
* Add workflowInstanceId and taskInstanceId in MDC
(cherry picked from commit db595b3eff)
* fix 10092: Task group does not take effect
* fix 10092: Task group does not take effect
* fix 10092: Task group does not take effect
(cherry picked from commit ee2b855ced)
* The selected task instance was recreated when the Master service failed over.
* Returns the expression result directly.
* Use Recovery to determine whether to use the old task instance.
(cherry picked from commit dbdbfeaeee)
* fix cmd param to overwrite global param when executing complement
* fix cmd param to overwrite global param when executing complement
(cherry picked from commit d4aeee16e5)
* After the serial wait execution strategy stops the running workflow instance, the instance will be woken up and executed if there is a wait instance.
* clear logic
* Resource overloading
* The thread cache task flow definition should get the latest version.
* Coverage on New Code
* Coverage on New Code
* Coverage on New Code
* use an existing method.
* Increase unit test coverage.
* Task flow definitions enforce policy isolation.
* [common] Make dolphinscheduler_env.sh work
* Change dist tarball `dolphinscheduler_env.sh` location
from `bin/` to `conf/`, which users could finish their
change configuration operation in one single directory.
and we only need to add `$DOLPHINSCHEDULER_HOME/conf`
when we start our sever instead of adding both
`$DOLPHINSCHEDULER_HOME/conf` and `$DOLPHINSCHEDULER_HOME/bin`
* Change the `start.sh`'s path of `dolphinscheduler_env.sh`
* Change the setting order of `dolphinscheduler_env.sh`
* `bin/env/dolphinscheduler_env.sh` will overwrite the `<server>/conf/dolphinscheduler_env.sh`
when start the server using `bin/dolphinsceduler_daemon.sh` or `bin/install.sh`
* Change the related docs
- add FailoverService.java
- move failover method from MasterRegistryClient to FailoverService
- move failover code from FailoverExecuteThread to FailoverService
This closes#9263
* [Fix-9065] [master] when task submit failed, remove from queue and add state event to thread
This closes#9065
* Update dolphinscheduler-master/src/main/java/org/apache/dolphinscheduler/server/master/runner/WorkflowExecuteThread.java
yes, your idea is better
Co-authored-by: caishunfeng <caishunfeng2021@gmail.com>
* [Bug-9065] [master] update logger info
Co-authored-by: guoshupei <guoshupei@lixiang.com>
Co-authored-by: caishunfeng <caishunfeng2021@gmail.com>
* rebase dev
* change task state to dispatch if dispatch success
* update task host when dispatch
* add dispatch task event
* test
* check tenant after enable auto create
* handle dispatch state
Co-authored-by: caishunfeng <534328519@qq.com>
Wenjun Ruan
WangJPLeo
devosend
Jiajie Zhong
旺阳
BaoLiang
xiangzihao
caishunfeng
Tq
Jiajie Zhong
Paul Zhang
gaojun2048
sparklezzz
czeming
worry
Eric Gao
guoshupei