Present: Arne, Daniel, Eric B., Eric S., Hervé, Luca A., Luca G., Nikos
Deploy image status: Burn-in
- current image branch has CPU & RAM
- network and disk to be done
- how to structure the burn-in tests
- individual tests for CPU, disk, RAM, network
- umbrella test that runs all tests (to be run first, then individual ones in case sth needs to be repeated)
- ACTION: Daniel to provide skeleton image
- ACTION: Luca/Nikos to provide actual tests
- ACTION: Arne to check with upstream on "hold on failure"
- ACTION: Arne to check with upstream on "pass with errors"
Deploy image status: benchmarking
- if possible, there should be only one approach to collect and store benchmarking data
- we'll need to compare CF benchmarking numbers with HEP/WLCG benchmarking
- ACTION: Domenico/Daniel/Arne to add benchmarking "clean" step (based on the continuous benchmarking)
- no real support in ironic
- need to identify machines which go into final decommissioning (autodecomm tool)
- best approach for now is probably to use "maintenance" in conjunction with cleaning
- ACTION: Arne to check for ironic support
- ACTION: Arne to propose workflow
There are minutes attached to this event.