Tom Scogland
Tom Scogland
That helps @grondo, thanks! From a logical perspective I can see it fitting in with those. That said, if we want to expose those to fluxion I would think attributes...
> Seems easy in principle, but I'm not sure about the auto-detection of Dockerfile changes. This was the other thing I was thinking about earlier. We don't currently do it...
The mpich thing is because Ubuntu pulled the mpich deb from Debian during one of their tests turning on PMIX mode, so the upstream MPICH on Noble can’t bootstrap without...
Interesting thought. Currently we only evaluate properties on nodes, but fluxion actually _stores_ properties on every vertex. In principle we could do this with relatively few changes (though non-negligible evaluation...
That's pretty much what I would have expected to happen. I'm sure we can fix this, but as it sits we don't just validate it, the queue is the actual...
They are not meant to be, but some of them operate on timeouts and can have failures like that especially if the resource is overloaded. The first of those tests...
Is it just the bootstrap test? If you have changes to the timeouts I wouldn't mind considering them upstream.
Apparently this is also still active.
@grondo, is this still something you want to consider? happened to pop up in a search today
We discussed this offline yesterday. The cause may be slot related, but I would expect this behavior even if it's just one slot, because the current implementation sets the max...