Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
apache
DolphinScheduler
提交
034475bf
DolphinScheduler
项目概览
apache
/
DolphinScheduler
上一次同步 接近 2 年
通知
707
Star
9572
Fork
3514
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
DolphinScheduler
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
提交
034475bf
编写于
7月 27, 2022
作者:
C
caishunfeng
提交者:
Jiajie Zhong
7月 28, 2022
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
[Bug-11101] fix task failover NPE (#11168)
(cherry picked from commit
6c7e00c4
)
上级
88712b42
变更
1
隐藏空白更改
内联
并排
Showing
1 changed file
with
16 addition
and
8 deletion
+16
-8
dolphinscheduler-master/src/main/java/org/apache/dolphinscheduler/server/master/service/MasterFailoverService.java
...cheduler/server/master/service/MasterFailoverService.java
+16
-8
未找到文件。
dolphinscheduler-master/src/main/java/org/apache/dolphinscheduler/server/master/service/MasterFailoverService.java
浏览文件 @
034475bf
...
...
@@ -40,6 +40,7 @@ import org.apache.dolphinscheduler.server.master.runner.task.TaskProcessorFactor
import
org.apache.dolphinscheduler.server.utils.ProcessUtils
;
import
org.apache.dolphinscheduler.service.process.ProcessService
;
import
org.apache.dolphinscheduler.service.registry.RegistryClient
;
import
org.apache.dolphinscheduler.spi.utils.StringUtils
;
import
org.apache.commons.collections4.CollectionUtils
;
import
org.apache.commons.lang3.time.StopWatch
;
...
...
@@ -232,14 +233,7 @@ public class MasterFailoverService {
// kill worker task, When the master failover and worker failover happened in the same time,
// the task may not be failover if we don't set NEED_FAULT_TOLERANCE.
// This can be improved if we can load all task when cache a workflowInstance in memory
try
{
TaskKillRequestCommand
killCommand
=
new
TaskKillRequestCommand
(
taskInstance
.
getId
());
Host
workerHost
=
Host
.
of
(
taskInstance
.
getHost
());
nettyExecutorManager
.
doExecute
(
workerHost
,
killCommand
.
convert2Command
());
LOGGER
.
info
(
"Failover task success, has killed the task in worker: {}"
,
taskInstance
.
getHost
());
}
catch
(
ExecuteException
e
)
{
LOGGER
.
error
(
"Kill task failed"
,
e
);
}
sendKillCommandToWorker
(
taskInstance
);
}
else
{
LOGGER
.
info
(
"The failover taskInstance is a master task"
);
}
...
...
@@ -249,6 +243,20 @@ public class MasterFailoverService {
processService
.
saveTaskInstance
(
taskInstance
);
}
private
void
sendKillCommandToWorker
(
@NonNull
TaskInstance
taskInstance
)
{
if
(
StringUtils
.
isEmpty
(
taskInstance
.
getHost
()))
{
return
;
}
try
{
TaskKillRequestCommand
killCommand
=
new
TaskKillRequestCommand
(
taskInstance
.
getId
());
Host
workerHost
=
Host
.
of
(
taskInstance
.
getHost
());
nettyExecutorManager
.
doExecute
(
workerHost
,
killCommand
.
convert2Command
());
LOGGER
.
info
(
"Failover task success, has killed the task in worker: {}"
,
taskInstance
.
getHost
());
}
catch
(
ExecuteException
e
)
{
LOGGER
.
error
(
"Kill task failed"
,
e
);
}
}
private
boolean
checkTaskInstanceNeedFailover
(
@NonNull
TaskInstance
taskInstance
)
{
if
(
taskInstance
.
getState
()
!=
null
&&
taskInstance
.
getState
().
typeIsFinished
())
{
// The task is already finished, so we don't need to failover this task instance
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录