Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-74783

Improve serverStatus test coverages

    XMLWordPrintableJSON

Details

    • Icon: Engineering Test Engineering Test
    • Resolution: Unresolved
    • Icon: Major - P3 Major - P3
    • None
    • None
    • None
    • None
    • Replication
    • Repl 2024-02-19

    Description

      Atlas consumes the serverStatus output for a lot of things (e.g. InTel). And even a small behavior change in serverStatus output could break a Cloud dashboard. Ideally, those should be caught by Cloud testing during the rc period. But the large surface area makes it difficult or infeasible. So maybe it should be a joint effort from both Cloud and Server. This ticket is to think about how to improve the current serverStatus test coverages.

      Most serverStatus tests I found in the Server repo are pretty basic and targeted. And it also seems impractical for the Server to test everything in all cases on a running MongoDB node. While I think it is difficult to test that everything is correct (i.e. every field reporting the exact value precisely), I think it might be easier if we just want to test everything has meaningful/valid value. Maybe we can explore extending fsm_workloads/server_status.js or adding a background hook to check the validity of serverStatus output while running the server test suites.

      Attachments

        Activity

          People

            pavithra.vetriselvan@mongodb.com Pavithra Vetriselvan
            lingzhi.deng@mongodb.com Lingzhi Deng
            Votes:
            0 Vote for this issue
            Watchers:
            7 Start watching this issue

            Dates

              Created:
              Updated: