2020-06-02 Smith Dotson QCFractal

Date

Jun 2, 2020

Participants

  • @David Dotson

  • @Daniel Smith (Deactivated)

Goals

  • I had a set of questions for Daniel:

    • anything special needed for set up of torsion drives locally?

      • service setup?

    • can we mark incompletes as errored?

    • can we mark errored reliably for re-run?

Discussion notes

  • Rerunning services will only see tasks have failed.

    • restart service after restarting tasks

  • For tasks that are incomplete, try restarting manager on offending resource

  • Would be useful to add a tag to a problem calculation and then only run those tags on a resource you control (such as a workstation)

    • need to pursue adding the ability to add tags in this way with Ben or Doaa

  • As for logging, can probably attach whole log file contents to stdout, stderr

  • May consider in the future standing up a logging system deployment that is queryable (ELK, sentry)

Action items

@David Dotson will pursue pathway for directing problem tasks to infrastructure he can introspect

Decisions