sanderegg
commented
3 years ago Update on sprint Red Panda
E2E
- added Jupyters e2e test running in parallel to simulate 20 users on master deploy, 10 users on staging AWS and production AWS deploys
- A number of issues were detected and fixed to improve the success rate of the e2e/p2p testing (see platform stability #1426 case)
Maintenance
- 3 data loss problems were detected:
- Jupyter lab state data was lost due to (un)archiving issue. Issue was fixed through #2192. Also DevOps change to ensure automated backups of S3 storage is done and also versioning which should prevent loosing data. Currently tested in on-premise master deploy, and will be moved to staging/production (on-premise and AWS) next sprint.
- A user opened a shared study and could not see the sharer user data in a Jupyter lab. After closing the study the empty data overwrote the original data, thus loosing the original Jupyter lab data. Issue was fixed through #2041.
- User A with a shared study opened got disconnected from the platform for a long time. User B worked on the shared study and closed it properly. User A eventually re-connected, this overwrote the data of user B. Issue was fixed through #2225
- All maintenance cases are listed in platform stability #1426. in the future will be connected to this issue.
NOTE: these fixes are currently all deployed in master/staging/staging_AWS. They will be deployed to production ASAP.
mguidon
elisabettai
the scrum master wrote: We are having quite a hard time maintaining the platformS, so that they pass our E2E tests [...]
in this sprint:
in the following ones: