* Refactor code to support tracing
* Extension network protocol, support context and version
* Extension master asynchronous queue support context
* Extract scan task method from MasterSchedulerService for tracing
* fix
* fix
* add test case
* fix
* fix
Co-authored-by: hailin0 <hailin0@yeah.net>
* [Fix-4268] Fix NumberFormatException when visiting a doc.html or swagger-ui.html
* Fix checkstyle error
* Update swagger-models in known-dependencies.txt
* remove getCurTaskParamsClass() in AbstractTask.java and replace it with TaskParametersUtils.getParameters()
* remove unused imports in AbstractTask.java
* reformat
* when the tenant does not exist, the task execution should throw an exception
* remote method createWorkDirAndUserIfAbsent
* set the task status failed when the tenant code does not exist.
* add taskLog.
* update check os user exists
* update TaskExecuteThreadTest test method.
* solving sonar fail.
Add no-var rule for eslint, and fix related files
Add npm run lint:fix command, which distinguished from the default command npm run lint
The behavior of command npm run lint and command npm run lint:fix should be distinguished
The first command is used to lint only, and the second one is used to lint and fix problems automatically
* fix(ui): re-login problem in new tab
* refresh page in new tab automatically
* support to reload router view in lightweight
* optimize visibility code
* [FIX-#3177]Task time parameter parsing error
rerun schedule time error
this closes#3177
* [FIX-#3177]Task time parameter parsing error
rerun schedule time error
this closes#3177
deleted the code at following two places:
org.apache.dolphinscheduler.service.process.ProcessService#checkTaskExistsInTaskQueue
org.apache.dolphinscheduler.service.process.ProcessService#taskZkInfo
* [FIX-#4172][server-worker] kill task NPE
The cache task will be sent when the Process is generated. Before that, if a kill task appears, then NPE will appear
Modification method: write into the cache when the task is received, and mark it as preData
If the task is killed before the Process is generated, delete the cache directly at this time
It will be judged before the process is generated. If the task has been killed, it will not be executed.
After the new process is created, write it into the cache, and judge again, if kill, then kill the process.
this closes#4172
* Delete the commented out code
Add spring beans
* code smell
* add test
* add test
* fix error
* test
* test
* revert
* fix error
* When there are tasks with assignment failure and the number of tasks in the current task queue is less than 10, sleep for 1 second
* When there are tasks with assignment failure and the number of tasks in the current task queue is less than 10, sleep for 1 second
* fix code smell & code style
* fix code smell & code style
Co-authored-by: zhanglong <zhanglong@ysstech.com>
* [Fix][Flink] fix flink args build problem
* [Fix][Flink] fix FlinkArgsUtilsTest
* [Improvement][UI] hide version and cluster input when deployMode is local
* fix 4190: When the amount of json data is large, process list page display slowly
* fix 4190: When the amount of json data is large, process list page display slowly
* fix 4190: When the amount of json data is large, process list page display slowly
Co-authored-by: baoliang <baoliang@analysys.com.cn>
When the message of successful execution arrives earlier than
the message of ack,
the message of ack will be discarded,
resulting in some information missing
break60
zh0122
hailin0
felix.wang
Shiwen Cheng
xingchun-chen
zhuangchong
Jatham
Tq
dailidong
lgcareer
CalvinKirs
Saksham Gupta
bao liang
BoYiZhang
chengp
wangxj3