Hi Vino,
Thank you for following up and creating the issue.
Best,
Gary
On Sun, Sep 9, 2018 at 10:02 AM, vino yang wrote:
> Hi Gary,
>
> Hi Gary, your guess about the scene is correct.
> We encountered this problem a month or two ago (sorry, there is no context
> log, but I think the problem is
Hi Gary,
Hi Gary, your guess about the scene is correct.
We encountered this problem a month or two ago (sorry, there is no context
log, but I think the problem is clear and not difficult to reproduce),
we will directly split it into trigger savepoint and cancel operation.
Devin worked with me at
Hi Devin,
If I understand you correctly, you are submitting a job in the YARN per-job
cluster mode. You are then invoking the "cancel with savepoint" command but
the client is not able to poll for the savepoint location before the cluster
shuts down.
I think your analysis is correct. As far as I
Hi Vino and Devin,
could you maybe send us the cluster entrypoint and client logs once you
observe the exception? That way it will be possible to debug it.
Cheers,
Till
On Tue, Sep 4, 2018 at 2:26 PM vino yang wrote:
> Hi Devin,
>
> Why do you trigger cancel with savepoint immediately after th
Hi Devin,
could you send the logs of the cluster entrypoint and the client once you
see this exception? This will help to debug the problem.
Cheers,
Till
On Tue, Sep 4, 2018 at 2:12 PM devinduan(段丁瑞) wrote:
> Hi all,
> I submit a flink job through yarn-cluster mode and cancel job with
>
Hi Devin,
Why do you trigger cancel with savepoint immediately after the job status
changes to Deployed? A more secure way is to wait for the job to become
running after it has been running for a while before triggering.
We have also encountered before, there will be a case where the client
times
Hi all,
I submit a flink job through yarn-cluster mode and cancel job with
savepoint option immediately after job status change to deployed. Sometimes i
met this error:
org.apache.flink.util.FlinkException: Could not cancel job .
at
org.apache.flink.client.cli.CliFrontend.lamb
Hi all,
I submit a flink job through yarn-cluster mode and cancel job with
savepoint option immediately after job status change to deployed. Sometimes i
met this error:
org.apache.flink.util.FlinkException: Could not cancel job .
at
org.apache.flink.client.cli.CliFrontend.lamb
