Rest API cancel-with-savepoint: 404s when passing path as target-directory

[Emily McMahon]

I've tried every combination I can think of to pass an s3 path as the target directory (url encode, include trailing slash, etc)

I can successfully pass a local path as the target directory (ie /jobs/$jobID/cancel-with-savepoint/target-directory/tmp) so I don't think there's a problem with the jobId or rest of the url. I also verified I can create the savepoint on s3 from the command line so it's not a permission issue.

Here's the same question on stack overflow (with the exception that they are getting a 502 whereas I'm getting a 404)

using Flink 1.3.1

Anyone have a working example?

Thanks,

Emily

[Fabian Hueske-2]

Hi Emily,

thanks for reaching out.

I'm not familiar with the details of the Rest API but Ufuk (in CC) might be able to help you.

Best, Fabian

2017-09-19 10:23 GMT+02:00 Emily McMahon <[hidden email]>:

I've tried every combination I can think of to pass an s3 path as the target directory (url encode, include trailing slash, etc)

I can successfully pass a local path as the target directory (ie /jobs/$jobID/cancel-with-savepoint/target-directory/tmp) so I don't think there's a problem with the jobId or rest of the url. I also verified I can create the savepoint on s3 from the command line so it's not a permission issue.

Here's the same question on stack overflow (with the exception that they are getting a 502 whereas I'm getting a 404)

using Flink 1.3.1

Anyone have a working example?

Thanks,

Emily

[Eron Wright]

Good news, it can be done if you carefully encode the target directory with percent-encoding, as per:

https://tools.ietf.org/html/rfc3986#section-2.1

For example, given the directory `s3:///savepoint-bucket/my-awesome-job`, which encodes to `s3%3A%2F%2F%2Fsavepoint-bucket%2Fmy-awesome-job`, I was able to submit the following URL:

http://localhost:8081/jobs/5c360ded6e4b7d8db103e71d68b7c83d/cancel-with-savepoint/target-directory/s3%3A%2F%2F%2Fsavepoint-bucket%2Fmy-awesome-job

And see the following in the log:

2017-09-19 14:27:45,939 INFO  org.apache.flink.runtime.jobmanager.JobManager                - Trying to cancel job 5c360ded6e4b7d8db103e71d68b7c83d with savepoint to s3:///savepoint-bucket/my-awesome-job

-Eron

On Tue, Sep 19, 2017 at 1:54 PM, Fabian Hueske <[hidden email]> wrote:

Hi Emily,

thanks for reaching out.

I'm not familiar with the details of the Rest API but Ufuk (in CC) might be able to help you.

Best, Fabian

2017-09-19 10:23 GMT+02:00 Emily McMahon <[hidden email]>:

I've tried every combination I can think of to pass an s3 path as the target directory (url encode, include trailing slash, etc)

I can successfully pass a local path as the target directory (ie /jobs/$jobID/cancel-with-savepoint/target-directory/tmp) so I don't think there's a problem with the jobId or rest of the url. I also verified I can create the savepoint on s3 from the command line so it's not a permission issue.

Here's the same question on stack overflow (with the exception that they are getting a 502 whereas I'm getting a 404)

using Flink 1.3.1

Anyone have a working example?

Thanks,

Emily

... [show rest of quote]

[Emily McMahon]

Thanks Eron & Fabian.

The issue was hitting a yarn proxy url vs the node itself. For example this worked 

http://{ip}:37716/jobs/1a0fd176ec8aabb9b8464fa481f755f0/cancel-with-savepoint/target-directory/s3%3A%2F%2F%2Fremit-flink 

But this did not

http://{ip}:20888/proxy/application_1504649135200_0001/jobs/1a0fd176ec8aabb9b8464fa481f755f0/cancel-with-savepoint/target-directory/s3%3A%2F%2F%2Fremit-flink

It's a bit confusing because the cancel api works with either and the proxy url sometimes works as this was successful http://{ip}:20888/proxy/application_1504649135200_0001/jobs/cca2dd609c716a7b0a195700777e5b1f/cancel-with-savepoint/target-directory/tmp/

Cheers,

Emily

On Tue, Sep 19, 2017 at 2:37 PM, Eron Wright <[hidden email]> wrote:

Good news, it can be done if you carefully encode the target directory with percent-encoding, as per:

https://tools.ietf.org/html/rfc3986#section-2.1

For example, given the directory `s3:///savepoint-bucket/my-awesome-job`, which encodes to `s3%3A%2F%2F%2Fsavepoint-bucket%2Fmy-awesome-job`, I was able to submit the following URL:

http://localhost:8081/jobs/5c360ded6e4b7d8db103e71d68b7c83d/cancel-with-savepoint/target-directory/s3%3A%2F%2F%2Fsavepoint-bucket%2Fmy-awesome-job

And see the following in the log:

2017-09-19 14:27:45,939 INFO  org.apache.flink.runtime.jobmanager.JobManager                - Trying to cancel job 5c360ded6e4b7d8db103e71d68b7c83d with savepoint to s3:///savepoint-bucket/my-awesome-job

-Eron

On Tue, Sep 19, 2017 at 1:54 PM, Fabian Hueske <[hidden email]> wrote:

Hi Emily,

thanks for reaching out.

I'm not familiar with the details of the Rest API but Ufuk (in CC) might be able to help you.

Best, Fabian

2017-09-19 10:23 GMT+02:00 Emily McMahon <[hidden email]>:

I've tried every combination I can think of to pass an s3 path as the target directory (url encode, include trailing slash, etc)

I can successfully pass a local path as the target directory (ie /jobs/$jobID/cancel-with-savepoint/target-directory/tmp) so I don't think there's a problem with the jobId or rest of the url. I also verified I can create the savepoint on s3 from the command line so it's not a permission issue.

Here's the same question on stack overflow (with the exception that they are getting a 502 whereas I'm getting a 404)

using Flink 1.3.1

Anyone have a working example?

Thanks,

Emily

... [show rest of quote]

... [show rest of quote]

[Nico Kruber]

Hi Emily,
I'm not familiar with the details of the REST API either but if this is a
problem with the proxy, maybe it is already interpreting the encoded URL and
passes it on un-encoded - have you tried encoding the path again? That is,
encoding the percent-signs:

http://
{ip}:20888/proxy/application_1504649135200_0001/jobs/
1a0fd176ec8aabb9b8464fa481f755f0/cancel-with-savepoint/target-directory/
s3%253A%252F%252F%252Fremit-flink


Nico

On Wednesday, 20 September 2017 00:52:05 CEST Emily McMahon wrote:

> Thanks Eron & Fabian.
>
> The issue was hitting a yarn proxy url vs the node itself. For example this
> worked
> http://
> {ip}:37716/jobs/1a0fd176ec8aabb9b8464fa481f755f0/cancel-with-savepoint/targe
> t-directory/s3%3A%2F%2F%2Fremit-flink
>
> But this did not
> http://
> {ip}:20888/proxy/application_1504649135200_0001/jobs/1a0fd176ec8aabb9b8464fa
> 481f755f0/cancel-with-savepoint/target-directory/s3%3A%2F%2F%2Fremit-flink
>
> It's a bit confusing because the cancel api works with either and the proxy
> url sometimes works as this was successful http://
> {ip}:20888/proxy/application_1504649135200_0001/jobs/cca2dd609c716a7b0a19570
> 0777e5b1f/cancel-with-savepoint/target-directory/tmp/
>
> Cheers,
> Emily
>
> On Tue, Sep 19, 2017 at 2:37 PM, Eron Wright <[hidden email]> wrote:
> > Good news, it can be done if you carefully encode the target directory
> > with percent-encoding, as per:
> > https://tools.ietf.org/html/rfc3986#section-2.1
> >
> > For example, given the directory `s3:///savepoint-bucket/my-awesome-job`,
> > which encodes to `s3%3A%2F%2F%2Fsavepoint-bucket%2Fmy-awesome-job`, I was
> > able to submit the following URL:
> > http://localhost:8081/jobs/5c360ded6e4b7d8db103e71d68b7c8
> > 3d/cancel-with-savepoint/target-directory/s3%3A%2F%2F%
> > 2Fsavepoint-bucket%2Fmy-awesome-job
> >
> > And see the following in the log:
> > 2017-09-19 14:27:45,939 INFO
> > org.apache.flink.runtime.jobmanager.JobManager>
> >                - Trying to cancel job 5c360ded6e4b7d8db103e71d68b7c83d
> >
> > with savepoint to s3:///savepoint-bucket/my-awesome-job
> >
> > -Eron
> >
> > On Tue, Sep 19, 2017 at 1:54 PM, Fabian Hueske <[hidden email]> wrote:
> >> Hi Emily,
> >>
> >> thanks for reaching out.
> >> I'm not familiar with the details of the Rest API but Ufuk (in CC) might
> >> be able to help you.
> >>
> >> Best, Fabian
> >>
> >> 2017-09-19 10:23 GMT+02:00 Emily McMahon <[hidden email]>:
> >>> I've tried every combination I can think of to pass an s3 path as the
> >>> target directory (url encode, include trailing slash, etc)
> >>>
> >>> I can successfully pass a local path as the target directory (ie
> >>> /jobs/$jobID/cancel-with-savepoint/target-directory/tmp) so I don't
> >>> think there's a problem with the jobId or rest of the url. I also
> >>> verified
> >>> I can create the savepoint on s3 from the command line so it's not a
> >>> permission issue.
> >>>
> >>> Here's the same question on stack overflow
> >>> <https://stackoverflow.com/questions/45844298/send-flink-job-manager-a-f
> >>> ully-qualified-path-to-a-savepoint-directory-using-the> (with the
> >>> exception that they are getting a 502 whereas I'm getting a 404)
> >>>
> >>> using Flink 1.3.1
> >>>
> >>> Anyone have a working example?
> >>>
> >>> Thanks,
> >>> Emily

... [show rest of quote]

... [show rest of quote]

[Eron Wright]

It is not surprising to see fidelity issues with the YARN proxy.  I suggest opening a ticket on Flink side to update the cancel-with-savepoint API to take the target directory as a query string parameter (of course, backwards compatibility should be maintained).

On Wed, Sep 20, 2017 at 1:55 AM, Nico Kruber <[hidden email]> wrote:

Hi Emily,
I'm not familiar with the details of the REST API either but if this is a
problem with the proxy, maybe it is already interpreting the encoded URL and
passes it on un-encoded - have you tried encoding the path again? That is,
encoding the percent-signs:

http://
{ip}:20888/proxy/application_1504649135200_0001/jobs/
1a0fd176ec8aabb9b8464fa481f755f0/cancel-with-savepoint/target-directory/
s3%253A%252F%252F%252Fremit-flink


Nico

On Wednesday, 20 September 2017 00:52:05 CEST Emily McMahon wrote:
> Thanks Eron & Fabian.
>
> The issue was hitting a yarn proxy url vs the node itself. For example this
> worked
> http://
> {ip}:37716/jobs/1a0fd176ec8aabb9b8464fa481f755f0/cancel-with-savepoint/targe
> t-directory/s3%3A%2F%2F%2Fremit-flink
>
> But this did not
> http://
> {ip}:20888/proxy/application_1504649135200_0001/jobs/1a0fd176ec8aabb9b8464fa
> 481f755f0/cancel-with-savepoint/target-directory/s3%3A%2F%2F%2Fremit-flink
>
> It's a bit confusing because the cancel api works with either and the proxy
> url sometimes works as this was successful http://
> {ip}:20888/proxy/application_1504649135200_0001/jobs/cca2dd609c716a7b0a19570
> 0777e5b1f/cancel-with-savepoint/target-directory/tmp/
>
> Cheers,
> Emily
>
> On Tue, Sep 19, 2017 at 2:37 PM, Eron Wright <[hidden email]> wrote:
> > Good news, it can be done if you carefully encode the target directory
> > with percent-encoding, as per:
> > https://tools.ietf.org/html/rfc3986#section-2.1
> >
> > For example, given the directory `s3:///savepoint-bucket/my-awesome-job`,
> > which encodes to `s3%3A%2F%2F%2Fsavepoint-bucket%2Fmy-awesome-job`, I was
> > able to submit the following URL:
> > http://localhost:8081/jobs/5c360ded6e4b7d8db103e71d68b7c8
> > 3d/cancel-with-savepoint/target-directory/s3%3A%2F%2F%
> > 2Fsavepoint-bucket%2Fmy-awesome-job
> >
> > And see the following in the log:
> > 2017-09-19 14:27:45,939 INFO
> > org.apache.flink.runtime.jobmanager.JobManager>
> >                - Trying to cancel job 5c360ded6e4b7d8db103e71d68b7c83d
> >
> > with savepoint to s3:///savepoint-bucket/my-awesome-job
> >
> > -Eron
> >
> > On Tue, Sep 19, 2017 at 1:54 PM, Fabian Hueske <[hidden email]> wrote:
> >> Hi Emily,
> >>
> >> thanks for reaching out.
> >> I'm not familiar with the details of the Rest API but Ufuk (in CC) might
> >> be able to help you.
> >>
> >> Best, Fabian
> >>
> >> 2017-09-19 10:23 GMT+02:00 Emily McMahon <[hidden email]>:
> >>> I've tried every combination I can think of to pass an s3 path as the
> >>> target directory (url encode, include trailing slash, etc)
> >>>
> >>> I can successfully pass a local path as the target directory (ie
> >>> /jobs/$jobID/cancel-with-savepoint/target-directory/tmp) so I don't
> >>> think there's a problem with the jobId or rest of the url. I also
> >>> verified
> >>> I can create the savepoint on s3 from the command line so it's not a
> >>> permission issue.
> >>>
> >>> Here's the same question on stack overflow

> >>> <https://stackoverflow.com/questions/45844298/send-flink-job-manager-a-f
> >>> ully-qualified-path-to-a-savepoint-directory-using-the> (with the

> >>> exception that they are getting a 502 whereas I'm getting a 404)
> >>>
> >>> using Flink 1.3.1
> >>>
> >>> Anyone have a working example?
> >>>
> >>> Thanks,
> >>> Emily

... [show rest of quote]