go-deploy | Support svn | Continuous Deployment library

It looks like you need to change the config of your Springboot app to use the mongodb endpoint: mongodb:27017 rather than localhost:27017❓🤔. Since mongo is not running on the same container/pod anymore.

In the application.properties file something like this:

The double dash symbol "--" is used to separate the arguments you want to pass to the command from the kubectl arguments. So the correct way is:

Well to answer your question, there are few things that you have to check out while deploying your Django or flask or dash application using AWS ElasticBeanstalk(EBS).

NOTE: The following steps are outlined if the application is hosted using AWS EBSCLI i.e the command line feature of EBS and not the method of hosting through the User Interface of the EBS on your AWS account. On my personal experience, I have found that the UI method of hosting the application will not work for flask, Django or dash applications.

1. While hosting your application into EBS, check the hosting logs if you have got any errors. Even though the EBS prompts you that you have hosted your application successfully, you will have to cross check on the logs to find any errors.
2. Once you have verified you have no errors in Step 1, navigate to the logs console of your EBS application, download entire logs and then check for any other errors. If you have any errors, you have to rectify them first.
3. Now that you have said that the AWS Linux machine is not able to import the Django package, it might be that you are checking for the Django package manually under a different virtual environment whereas the AWS EBS has hosted your application under another virtual environment.
4. To solve Step 3, it happens to be that any application that is hosted using AWS EBS on a linux server is hosted under the following virtual environment path. /opt/python/run/venv/bin. In order to verify this, just type the command cd /opt/python/run/venv/bin after connecting to your linux server. Once you're able to navigate to this path, try activating the virtualenv by giving source ./activate
5. Once you've activated this virtualenv that AWS EBS has created for your application, try starting python inside this virtualenv and then try import django. If you find that django package is not installed under this virtualenv, you have to install it manually, since this is the virtualenv that your hosted application will refer to. And any new packages that need to be referenced by your hosted application will have to be installed under this virtualenv only.
6. To further give you additional information, your hosted application will reside on the AWS linux server under this directory /opt/python/bundle/2/app. You can refer to this directory if you are willing to alter your application code or any supporting files in the future. For example, if you want to change any of the source code, you can directly edit the source code files under this repository and restart your server for the changes to take effect.

Update: Since you've specified that during the hosting of your application, you have received the error on the django package not found, please follow the following steps to rectify this error.

1. While creating your requirements.txt file, make sure that only the necessary packages for your application are present inside that file. When you do pip freeze requirements.txt, all the packages which are present in your current environment will make an entry in requirements.txt file. This might create errors while AWS EBS is trying to install this file. The reason is, say for example, PackageA and PackageB might be dependent packages of PackageC. If you install PackageC, automatically PackageA and PackageB would get installed. Having said this, you have to remove PackageA and PackageB in your requirements.txt file and keep only PackageC. Because installation of PackageC would automatically install it's dependent packages which are PackageA and PackageB. So to keep it short, you only have to include the packages that you manually installed using pip install in your local machine inside your requirements.txt file. All other packages will be dependent packages of these packages and will install automatically when the master package is installed. So these dependent packages need to be removed from requirements.txt.
2. If you're still receiving the same error, once the application is hosted with errors, try to connect to your AWS Linux machine using CLI(Command Line Interface), navigate to the virtual environment path by doing cd /opt/python/run/venv/bin, activate it using source ./activate, invoke python interpreter and then try importing the packages manually. If you find any of the packages are uninstalled, you can install them manually and restart the server for the changes to take effect.

For more details, you can refer this link. Though this is created for hosting a flask application, hosting a django application will also work on same lines.

You have a typo in your mongodb service definition.

Use mongodb:// in front of your panel-admin-mongo-cluster-ip-service

So it should look like this: mongodb://panel-admin-mongo-cluster-ip-service

Yes it is possible with the reclaim setting. Please refer this documentation

[Edit]

Sorry my bad, the connections string mongodb://mongodb:27017 would actually work. I tried dns querying that name, and it was able to resolve to the correct ip address even without specifying ".default.svc...".

root@web-controller-mlplb:/app# host mongodb mongodb.default.svc.cluster.local has address 10.108.119.125

@Anshul Jindal is correct that you have race condition, where the web pods are being loaded first before the database pods. You were probably doing kubectl apply -f . Try doing a reset kubectl delete -f . in the folder containing those yaml . Then kubectl apply the database manifests first, then after a few seconds, kubectl apply the web manifests. You could also probably use Init Containers to check when the mongo service is ready, before running the pods. Or, you can also do that check in your node.js application.

Example of waiting for mongodb service in Node.js

In your connection.js file, you can change the connect function such that if it fails the first time (i.e due to mongodb service/pod not being available yet), it will retry again every 3 seconds until a connection can be established. This way, you don't even have to worry about load order of applying kubernetes manifests, you can just kubectl apply -f .

Turns out the issue was with this line in nginx config: