Everything under source (src) in backend was written except for the final test files and API was also complete. We were waiting for more information about access tokens and Gitlab access keys from the IAM team. We were also waiting to hear from AWS team on templates for Kubernetes system but did not get any response. We were at the same time researching on how to import API bundle into backend so we could start building the backend server and to at least test if everything from our side was working properly.
We needed to bundle our API before we could export the file to backend and ‘npm run bundle outfile.yaml` was not working. After fixing minor issues in the API, we got our bundle file. Even after a lot of research, readings and experimentation, we were not able to connect API bundle to backend, something would not work and kept getting different errors. At this point we decided to copy the bundle file and paste it into backend lib folder manually. Since we were not able to import the bundle file, we decided to not work in main and created a separate branch called ‘issue10’.
We made a lot of mistakes from this point on. We were all working on this branch instead of making our own branch based on the issue10 branch. We also worked only on one team member – Christian’s system. We added and modified a lot files like entrypoint.sh, package.json, package-lock.json, docker-compose.yaml, Dockerfile, etc. We should have created branches based on issue10 branch, then created specific issues with proper weight assigned to them. After making all the modifications we could think so, we still were not able to build our backend server.
At this point, Dr. Wurst pointed out that we do not have frontend, which means that our docker file will not be copying yarn file and running yarn install but instead copying the package-lock file and running npm install. After making changes in the Dockerfile, our backend server was finally able to build but still could not stay up. After going over docker container logs we finally came to the conclusion that the problem for this was in the port. We had not specified proper port in many files. Once those modifications were made the backend server was up. Till this point, we were all still working on Christian’s system, without any proper issues, comments and documentation for modifications that were being made.
I created my own branch based on issue10 and then wrote calls for cooking method. However, when I tried to build the server, I got the error that docker-compose is not installed even though I have used it many times before. Using the solution, I found on stackoverflow I was now able to build the server but the process would hang up on the attaching stage. I tried running docker-compose up command after opening the branch in container. This led to an error saying ‘docker not found’. Another team member Jeffery had the same problem as mine but was able to find a solution for it and then was able to make calls. However, the same solution did not work for me. After trying everything I could find on the web till 4 am, I finally gave up.
In the next meeting I shared the calls I had written for cooking methods with the team so they could be added to the project. Noticing that I was the only person who was working on a windows PC I decided to run the project on a friends MacBook and surprisingly it worked. I did not research further into why it would not work on my system since I had limited time with the Mac.
I worked on cookingMethod.js and cookingTip.js test file. I set three constant variables: getAll, getOne with valid ID and an invalid ID. getAll would check for ‘200’ response and data in form of array. getOne with valid input would check for ‘200’ response, ID, object and value. getOne with Invalid input would check for ‘404’ response.
Links to issues:
Modification of files
Cooking Method test
Cooking tip test
P.S. – I apologize for this retrospective being more like a summary. It proved difficult to classify the process in what worked and what did not work sections for this sprint.