action #53588
closed
o3: yaml output of tumbleweed group looks wrong
Added by coolo over 5 years ago.
Updated over 5 years ago.
Category:
Regressions/Crashes
Description
https://openqa.opensuse.org/admin/job_templates/1 creates this yaml:
...
- RAID0
- RAID0:
machine: 64bit
- RAID1:
machine: 64bit
priority: 51
- RAID1:
priority: 51
- RAID5:
priority: 51
- RAID5:
machine: 64bit
priority: 51
- RAID10:
priority: 51
- RAID10:
machine: 64bit
priority: 51
...
i.e. all RAIDs are duplicated
And this seems to be going through, e.g. in group 2 I have:
opensuse-*-Staging-DVD-x86_64:
- textmode
- kde:
machine: USBboot_64
priority: 45
- kde:
priority: 45
- gnome:
priority: 45
- minimalx:
priority: 45
- minimalx:
machine: uefi-staging
priority: 45
- RAID0
- RAID1
- cryptlvm
- rescue_system
- update_staging
- autoyast_mini
The title of this issue seems quite appropriate. The output looks wrong until you realize that the "duplicates" are run on multiple machines, ie. 32bit and 64bit. When defaults are applied the suite with the default machine will not show what machine it is.
@cdywan will you pick up this ticket or what do you suggest to do? I am not sure if it makes sense for anyone else to pick it up
I've just had a look at this, too. The "duplicated" RAIDs are actually for different machines:
- RAID1: // machine is 64bit
machine: 64bit
priority: 51
- RAID1: // machine is 32bit
priority: 51
In the 2nd RAID1 the 32bit has been omitted because it is the default, see the defaults defined at the top of the YAML:
defaults:
i586:
machine: 32bit
priority: 50
i586-x86_64:
machine: 32bit
priority: 50
i686:
machine: 32bit
priority: 50
x86_64:
machine: 64bit
priority: 50
QAM could reuse aliases between YAMLs within Update|Incidents groups e.g. there are subgroups which could use YAML alias of tests running on all versions, another alias running only on sle15 and so on.
But that sounds like a layer above this. I.e. maintain your input in a git repository, expand the yaml to your linking and post it to the API for all groups. But expecting the UI to support aliases and includes is not what I think we should do.
^ I copied the comments to #54179 - I think these ended up here by accident. This ticket is about defaults and perceived duplicates.
As discussed in the weekly this morning, I'm going to look into adapting the generated YAML. So we can try and avoid the "duplicates" when there's multiple tests with different machines.
- Status changed from New to Resolved
Also available in: Atom
PDF