Cannot see the results on instance while creation of new app.

[C4rohan]

I am trying to create a new aws serverless on the app Here is the aws serverless app link I uploaded to git. --> Link Here is the docker file for the app --> Link Here is the config files in examples --> Link

I am successful in running the the app. There are no errors while resource creation and the output is similar to the author's output.

The problem is I am unaware of ./AwsServerlessTemplate/NewAppTemplate/lambda and its working. I somehow managed to figure out the settings by looking into other applications.

Steps taken to execute the application

• create application.ini, resource.ini, personal.ini
• Author doesn't mention about deploy_config.json and SampleEvent.json but got many errors based on these two and eventually had to edit those as well. Mostly the issues were related to set settings for deploy_config.json indirectly indicating someissues with wrong configuration of aws.
• Here are some git issues
  1. Link
  2. Link

Output Observations and questions-->

• one click command is supposed to close the cloud formation, instead it takes forever to delete the cloud formation.

• I see that the instance that has been created is empty and when opened and monitored we see no execution of the [command ] or [bootstrap].

• I think there should be step by step explanation of how does someone decide or set the parameters for deploy_config.json and SampleEvent.json.

• Are the steps mentioned on the git the exact steps since nothing is mentioned about deploy_config.json and SampleEvent.json ?and eventually I had edit these two .

• Please let me know if any other information is needed on this matter.

[C4rohan]

The docker and the commands are tested on a testing instance and running correctly.. The current problem I believe is the same the commands and bootstrap I load doesnt execute on the app created instance. Main reason I think is because of aws set up of resources linked with deploy_config.json .

[starlyxxx]

Hi @C4rohan, thanks for your effort on this repository. Based on your screenshot, I found the CloudWatch Log on link. The log shows that the serverless pipeline cannot find the masterInstanceId.

In deploy_config.json, you have changed the parameter "DLSecurityGroup" to "sg-0f81c232cd437d43f". However, in app.py, I hardcoded the "DLSecurityGroup" to "distributed_dl_starly". I believe that editing the app.py file is a correct direction to solve all the issues.

Also, please try to use CloudWatch to debug the commands that run on the app created instance. It is helpful when I create every new aws serverless on the app.

[C4rohan]

Thank you for helping out for debugging .. I have changed the "DLSecurityGroup" to "sg-0f81c232cd437d43f" . Thank you again for letting me know about cloudwatch

[jianwuwang]

Thanks for working on it. I want to add some background. The current example only needs one GPU node. @C4rohan is trying to run Python code in serial with one GPU or in parallel with multiple GPUs on the same node.

[C4rohan]

@starlyxxx We have fallen into another error here is the link --> https://github.com/big-data-lab-umbc/Reproducible_and_portable_app_in_cloud/issues/14