Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Support for single test result appearing in multiple meshes #147

Open
timchown opened this issue Dec 11, 2024 · 3 comments
Open

Support for single test result appearing in multiple meshes #147

timchown opened this issue Dec 11, 2024 · 3 comments

Comments

@timchown
Copy link

It seems at the moment we can't use one test result, like a v6 throughout test, in multiple meshes, rather than having the test repeated because a pair of nodes are in two meshes.

Chris Walker asked about de-duplication of tests from multiple meshes in one of our monthly WLCG-dev calls and we recall the answer was that at present the tests would be duplicated, because the hashing takes into account the mesh name.

Is there a way we can have this supported, so we could have sites that are in multiple meshes not repeat tests?

@mfeit-internet2
Copy link
Member

The best way I can think of to solve this would be putting a deduplicate parameter in tasks that would allow those with identical tests to be combined at the shortest of the repeat intervals. Tasks using repeat-cron might have to be exempt from that as they'd be difficult to merge in a way that satisfies everyone. Archives would have to be combined so everyone who's expecting a result gets one.

@arlake228
Copy link
Contributor

You'd also have to combine reference/meta information which contains details about the grids it belongs to and some other info. This is really the biggest challenge. Its easy to compare test parameters of things you've already built, but when things like archives and descriptive information differ we don't have a way to merge those. pSConfig would essentially have to do a pass through all the configs once to get the list of tests and do the merging, likely holding all of that in memory. Once it had gone through everything it could start talking to pscheduler. This is a reasonable request, but its not a small feature add so would need to be part of a bigger release.

@timchown
Copy link
Author

ok, thanks. it's not super critical to have but it would certainly be handy for us in the future, where we have more than one site in multiple communities. I appreciate a pscheduler function to recognise a test pair is already scheduled and not reschedule more tests may be trickier than you might expect. OTOH if duplicate throughput tests are scheduled it would be nice to show all the extra test results between the same pair. Is it easier if everything shares the same archive server?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
Status: Ready
Development

No branches or pull requests

3 participants