Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-83360

Reconsider meaning/behavior of "no trace" fatal assertions

    XMLWordPrintableJSON

Details

    • Icon: Improvement Improvement
    • Resolution: Won't Fix
    • Icon: Major - P3 Major - P3
    • None
    • None
    • None
    • None
    • Service Arch
    • 135

    Description

      Fatal assertions have both a vanilla version and a "no trace" version. The former uses std::abort while the latter uses quickExit with ExitCode::abrupt. In some cases it does make sense to not log a backtrace, for instance if the user starts up the server on unsupported data files. However we also won't capture a core dump in this case, even if the system is set up to capture them upon unclean shutdown. This can make diagnosis difficult if the conditions leading to the (no trace) fatal assertion were unexpected.

      Attachments

        Activity

          People

            backlog-server-servicearch Backlog - Service Architecture
            gregory.noma@mongodb.com Gregory Noma
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: