Regarding

[Rashmi Varma]

Hello,

I am trying to setup flink project on my local machine and I have been trying to run flink example codes but I keep getting 'NoResourceAvailablleException".

Please find the stack trace here : http://pastebin.com/NgPfnQck

--

[Márton Balassi]

Hi Rashmi,

The issue is that although you have a JobManager running (the master component of a Flink cluster scheduling the jobs) there are no TaskManagers running (the components doing the actual work). Hence you got the log line "Resources available to scheduler: Number of instances=0, total number of slots=0, available slots=0". 

My bet without further information would be that you have created this cluster with bin/start-cluster.sh from the Flink binaries and the TaskManager did not start up. You can check both the TaskManager and JobManager logs in the log/ directory if you want to investigate the issue. Also it is worth to execute "jps" which lists the currently running Java processes and their names to see whether you have both.

My suggestion for a local setup is to stop the cluster you are running currently (bin/stop-cluster.sh) and start a local one instead with bin/start-local.sh. This starts up the JobManager and a single TaskManager in the same JVM. On a side note you can also run Flink programs from your IDE with simply hitting Run, there is no need to fire up a cluster manually unless you are testing the cluster itself.

Which cluster setup are you using?

Best,

Marton

On Sat, Oct 8, 2016 at 9:24 PM, Rashmi Varma <[hidden email]> wrote:

Hello,

I am trying to setup flink project on my local machine and I have been trying to run flink example codes but I keep getting 'NoResourceAvailablleException".

Please find the stack trace here : http://pastebin.com/NgPfnQck

--

[Márton Balassi]

I think Flink supports running on Windows, but developing against Flink on Windows is something that we do not test. I have to admit that my knowledge in that area is very limited, Fabian (ccd) is the goto person with that topic.

Could you download the Flink binaries and try running the examples [1] there to see whether that works?

If you have access to a Unix (virtual) machine could you try running from an IDE there? 

[1] https://ci.apache.org/projects/flink/flink-docs-release-1.1/quickstart/setup_quickstart.html#start-a-local-flink-cluster

On Sat, Oct 8, 2016 at 11:45 PM, Rashmi Varma <[hidden email]> wrote:

Hi,

Yes , I am getting those errors on Netbeans even for a basic WordCount implementation.Also, yes, I am on windows. 

Thanks,

Rashmi

On Sat, Oct 8, 2016 at 11:36 PM, Márton Balassi <[hidden email]> wrote:

Hey,

So you get those error when running directly from Netbeans? You mentioned using the bat file - are you on Windows then? :)

If you are running directly from an IDE you do not need to run a cluster in the background, Flink will create one on the fly for you.

On Sat, Oct 8, 2016 at 11:25 PM, Rashmi Varma <[hidden email]> wrote:

Hello Marton,

Thanks for responding. I am not using anything for the cluster at the moment. I started only local server with start-local.bat.And I am executing my Java programs in Netbeans with simple 'Run'. 

I have added all maven dependencies.Although I have one question , if I have already added flink dependencies, do I still need to start flink locally for task manager and job manager to run in the background.

Regards,

Rashmi

On Sat, Oct 8, 2016 at 11:08 PM, Márton Balassi <[hidden email]> wrote:

Hi Rashmi,

The issue is that although you have a JobManager running (the master component of a Flink cluster scheduling the jobs) there are no TaskManagers running (the components doing the actual work). Hence you got the log line "Resources available to scheduler: Number of instances=0, total number of slots=0, available slots=0". 

My bet without further information would be that you have created this cluster with bin/start-cluster.sh from the Flink binaries and the TaskManager did not start up. You can check both the TaskManager and JobManager logs in the log/ directory if you want to investigate the issue. Also it is worth to execute "jps" which lists the currently running Java processes and their names to see whether you have both.

My suggestion for a local setup is to stop the cluster you are running currently (bin/stop-cluster.sh) and start a local one instead with bin/start-local.sh. This starts up the JobManager and a single TaskManager in the same JVM. On a side note you can also run Flink programs from your IDE with simply hitting Run, there is no need to fire up a cluster manually unless you are testing the cluster itself.

Which cluster setup are you using?

Best,

Marton

On Sat, Oct 8, 2016 at 9:24 PM, Rashmi Varma <[hidden email]> wrote:

Hello,

I am trying to setup flink project on my local machine and I have been trying to run flink example codes but I keep getting 'NoResourceAvailablleException".

Please find the stack trace here : http://pastebin.com/NgPfnQck

--

--

--

[Fabian Hueske-2]

Hi Rashmi,

as Marton said, you do not need to start a local Flink instance (start-lcoal.bat) if you want to run programs from your IDE.
Maybe running a local instance causes a conflict when starting an instance from IDE.

Developing and running Flink programs on Windows should work, both from the IDE and using a local Flink instance.

Best, Fabian

2016-10-09 0:05 GMT+02:00 Márton Balassi <[hidden email]>:

I think Flink supports running on Windows, but developing against Flink on Windows is something that we do not test. I have to admit that my knowledge in that area is very limited, Fabian (ccd) is the goto person with that topic.

Could you download the Flink binaries and try running the examples [1] there to see whether that works?

If you have access to a Unix (virtual) machine could you try running from an IDE there? 

[1] https://ci.apache.org/projects/flink/flink-docs-release-1.1/quickstart/setup_quickstart.html#start-a-local-flink-cluster

On Sat, Oct 8, 2016 at 11:45 PM, Rashmi Varma <[hidden email]> wrote:

Hi,

Yes , I am getting those errors on Netbeans even for a basic WordCount implementation.Also, yes, I am on windows. 

Thanks,

Rashmi

On Sat, Oct 8, 2016 at 11:36 PM, Márton Balassi <[hidden email]> wrote:

Hey,

So you get those error when running directly from Netbeans? You mentioned using the bat file - are you on Windows then? :)

If you are running directly from an IDE you do not need to run a cluster in the background, Flink will create one on the fly for you.

On Sat, Oct 8, 2016 at 11:25 PM, Rashmi Varma <[hidden email]> wrote:

Hello Marton,

Thanks for responding. I am not using anything for the cluster at the moment. I started only local server with start-local.bat.And I am executing my Java programs in Netbeans with simple 'Run'. 

I have added all maven dependencies.Although I have one question , if I have already added flink dependencies, do I still need to start flink locally for task manager and job manager to run in the background.

Regards,

Rashmi

On Sat, Oct 8, 2016 at 11:08 PM, Márton Balassi <[hidden email]> wrote:

Hi Rashmi,

The issue is that although you have a JobManager running (the master component of a Flink cluster scheduling the jobs) there are no TaskManagers running (the components doing the actual work). Hence you got the log line "Resources available to scheduler: Number of instances=0, total number of slots=0, available slots=0". 

My bet without further information would be that you have created this cluster with bin/start-cluster.sh from the Flink binaries and the TaskManager did not start up. You can check both the TaskManager and JobManager logs in the log/ directory if you want to investigate the issue. Also it is worth to execute "jps" which lists the currently running Java processes and their names to see whether you have both.

My suggestion for a local setup is to stop the cluster you are running currently (bin/stop-cluster.sh) and start a local one instead with bin/start-local.sh. This starts up the JobManager and a single TaskManager in the same JVM. On a side note you can also run Flink programs from your IDE with simply hitting Run, there is no need to fire up a cluster manually unless you are testing the cluster itself.

Which cluster setup are you using?

Best,

Marton

On Sat, Oct 8, 2016 at 9:24 PM, Rashmi Varma <[hidden email]> wrote:

Hello,

I am trying to setup flink project on my local machine and I have been trying to run flink example codes but I keep getting 'NoResourceAvailablleException".

Please find the stack trace here : http://pastebin.com/NgPfnQck

--

--

--