search

galasa-dev / projectmanagement

Project Management repo for Issues and ZenHub
7 stars 4 forks source link

Requests to RSE API in all tests are failing #2019

Closed jadecarino closed 1 month ago

jadecarino commented 1 month ago

Describe the bug

Requests to the RSE API are now failing with a Connection refused exception since today.

Fails in method submitJCLNoSteps after sending a POST request to v1/jobs/string/:

Fails in method basicKSDSDefine after sending a GET request to v1/datasets:

Fails in method unixFileCreate after sending a GET request to v1/unixfiles:

Fails in method testExistingDS after sending a GET request to v1/datasets:

Steps to reproduce

  1. Run any of the above mentioned tests with galasactl
  2. Go to the IVTs run.log and see the failures

Expected behavior

We should be able to make requests to the RSE API.

Additional context

If applicable, add screenshots or any relevant output logs to help explain your problem.

jadecarino commented 1 month ago

I have created a new OneTicket (189801) for the DevIT team to look into in case there is an issue with the RSE API server we point these tests at.

I have also looked at the RSE API documentation which looks like the endpoints we are using are correct (apart from some trailing slashes but I doubt this would have suddenly become an issue). The API is still at v1 so that matches with our code.

jadecarino commented 1 month ago

Response on the OneTicket says asked me to retry the tests as a settings change was made yesterday to remediate a vulnerability. The change was rolled back temporarily to test if this had caused our failures. Rerunning the tests then passed (and original issue seen in story #1961 was observed again). OneTicket has advised we need to look at getting Galasa updated to connect to the RSE API in a more secure fashion to work with their vulnerability update.

jadecarino commented 1 month ago

Settings change has been rolled back, so no longer getting a 'Connection refused' exception. Same 'Read timed out exception' still occurring so closing this but keeping open the original defect #1961