Yes, it turned out that we only configured TM<->JM communication correctly
- inter TM config was missing, hence being "reject all". Thanks for the
suggestion!
On Mon, Aug 23, 2021 at 5:29 PM Arvid Heise wrote:
> It rather looks to me as if the task manager can not communicate with each
> other.
It rather looks to me as if the task manager can not communicate with each
other. Can you check your network policies? Are they allowed to communicate
on random ports?
On Mon, Aug 23, 2021 at 8:37 AM Zbyszko Papierski
wrote:
> Hi,
>
> No, they don't - only the job is being restarted after that,
Hi,
No, they don't - only the job is being restarted after that, without any
luck. Exception I provided is added to a exceptions list of the job itself.
On Mon, Aug 23, 2021 at 4:50 AM Caizhi Weng wrote:
> Hi!
>
> This might be that some task managers cannot reach out to the job manager
> in ti
Hi!
This might be that some task managers cannot reach out to the job manager
in time. Has any of the task manager instance restarted after this failure?
If yes, what does the log (Flink log and kubernetes log) of the failed task
manager say?
Zbyszko Papierski 于2021年8月20日周五 下午11:07写道：
> Hi!
>
>
