如果我们通过SavePoint取消工作,Job被取消了,SavePoint是失败如何现在还原这项工作
我已经使用本机Kubernetes部署并与SavePoint一起在应用程序模式下部署了Flink作业(我正在使用REST API命令),但是如果它未能完成保存点,则可以为该工作取消,因为它已经被取消了?
1) success:
{
"status":{"id":"COMPLETED"},
"operation":{
"location":"file:/savepoints-dir/savepoint-ad4025-dd46c1bd1c80"
}
}
2) failure:
{
"status":{"id":"COMPLETED"},
"operation":{
"failure-cause":{
"class": "java.util.concurrent.CompletionException",
"stack-trace": "..."
}
}
}
I have deployed flink job in application mode using native kubernetes deployment and stopping job along with savepoint (I'm using rest api command for that) but if it fails to complete savepoint is it possible to take savepoint for that job as it was cancelled already?
1) success:
{
"status":{"id":"COMPLETED"},
"operation":{
"location":"file:/savepoints-dir/savepoint-ad4025-dd46c1bd1c80"
}
}
2) failure:
{
"status":{"id":"COMPLETED"},
"operation":{
"failure-cause":{
"class": "java.util.concurrent.CompletionException",
"stack-trace": "..."
}
}
}
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
如果您仍然有一个检查站,则可以从中重新启动: https://nightlies.apache.org/flink/flink/flink/flink-docs-release-1.13/docs/ops/state/checkpoints/checkpoints/#resuming-from--a-.----- - 检查点
If you still have a checkpoint, you can restart from it: https://nightlies.apache.org/flink/flink-docs-release-1.13/docs/ops/state/checkpoints/#resuming-from-a-retained-checkpoint