2021-01-18 Meeting notes

• Andrew hasn't progressed the tracing code.
• The issue with the incorrect memberships (TNG-655) appears to have now resolved it self and jtrace now reflects the correct details. The process needs some detailed step by step documentation to help others to perform the task.
• Renalware haven't responded about the outstanding Trello items. George will chase and include Hugh Cairns in the email. One item that needs solving is the changes to the schema to correct missing date envelopes. Currently a work around is in place for RaDaR extracts but we need the ukrdc schema updates to correct this and then renalware will need to tweak their feed to match. Now that the ukrdc repository code is back in order Nick should be able to push his updates including this and then renalware can make use of this.
• George has released a new version of validatopn for testing. It had some issues building which were traced to missing pyqt elements and dlls in the build.
• Fiona has requested a backup of the alphastudy database as the ssh connection facility has not worked for trying to extract data directly. A file has been supplied but is not a data dump but the database structure. George suggested we just ask for an sql database backup which should provide all the information we require. Fiona will check what Ed Lamb plans for the site so we can confirm if we can shut down the site or if it needs to be kept online. We currently have no DSA or contract in place soo there may be some IG implications if we were to host the site etc.
• George has sent some details of the AKI issue to Andre on how the reloading of the AKI data can be done. Andre is looking at this and will finalise changes to the transplantfirst site this week.
•  Andrew has been trying to dockerise his development environment but has had some issues with deployment due to the need to use python2.7 to run platter to build the backend whilst everything else runs on python 3.7. Andrew will see if Joel's docker experience will help in identifying how to fix this. He has a working environment using docker-compose.
• Fiona has highlighted the following radar tickets RDR-741 and RDR-746 and Andrew is continuing to sort the issue with medications not showing due to the units. 

 George has had issues with the atlassian tools and believes the issue is related to the server memory. Most of the tools are running on the same server. Tim will try increasing the memory on the server. The applications all need to be updated as they haven't been updated for some time. We also need to look at whether switching to the cloud based solution would be preferable.

Joel highlighted the lack of useful Readme files on the repository. As a minimum the readme should have a link to the confluence documentation. Joel suggested that new repos could start with a readme that provides all the documentation and then this could be moved to confluence at release stage. The readme should still contain some general information about the codebase description of its basic use and purpose.

Joel also highlighted the large number of checkins to master rather than the use of branches. This made it a lot harder to understand the code bases and changes. It was agreed that branches should be used and Bitbucket options changed to prevent accidental master checkins. Another advantage of branches is that the tests can be run in a more controlled manner to check if the branch code has broken master etc. George highlighted this may have an issue with builds which also highlighted the fact that the build is a combined build and deploy in most cases. Joel recommended that we should have a separate deployment script to ensure we manage deployment but benefit from the automatic builds. It is clear we need to start re configuring the way we define builds and deployments using bamboo to make a more controlled and managed system.