Created attachment 1352088 [details] no cores used Description of problem: Version-Release number of selected component (if applicable): 5.9.0.8 How reproducible: Always Steps to Reproduce: 1. Add rhos cloud provider 2. Go to any instance's dashboard 3. Actual results: In container field there are no cores used for instance. Expected results: Additional info:
PR with fix for master is here: https://github.com/ManageIQ/manageiq-providers-openstack/pull/341
New commit detected on ManageIQ/manageiq-providers-openstack/master: https://github.com/ManageIQ/manageiq-providers-openstack/commit/4d0c9df33fedef10b29d9edcf8f9fa1a755c63f7 commit 4d0c9df33fedef10b29d9edcf8f9fa1a755c63f7 Author: Scott Seago <sseago> AuthorDate: Tue Aug 28 21:22:20 2018 -0400 Commit: Scott Seago <sseago> CommitDate: Tue Aug 28 21:22:20 2018 -0400 bug 1513086: add cores per socket to OpenStack cloud inventory parser https://bugzilla.redhat.com/show_bug.cgi?id=1513086 For OpenStack cloud, cpu_cores_per_socket was always set to 1 in the legacy refresh_parser since OpenStack flavors only have a single "number of CPUs" param. The new inventory parser was not setting this param at all, though, leaving it as nil. app/models/manageiq/providers/openstack/inventory/parser/cloud_manager.rb | 1 + 1 file changed, 1 insertion(+)
Still the same. I think this issue starts in Flavors. Flavors also have multiple CPUs and single core.
I'm not sure what you mean by "starts in Flavors". The flavors model doesn't even have an attribute for "cores per CPU" (it only has # of cpus), so the refresh code hard-codes this value to 1. When I reproduced this in my development environment, the problem I found was that this was only being done in the legacy refresh_parser. The new inventory parser wasn't setting this value to anything at all. If you're still seeing the problem, assuming your environment has my fix, I'm guessing that you're hitting another bug that for some reason isn't reproducing in my environment. In my env, without this fix I'm seeing the "no cores" problem but with the fix I don't see it. Can you confirm that you're testing in an env with this fix and, if so, could you provide a screenshot showing exactly what you're seeing now?
Okay, I added another screenshot as attachment. The only thing changed is that it has 1 core instead of 0 cores. When I provision an instance with 4 Cores it shows in its summary that it uses 4 sockets instead of 4 cores and 1 core instead of 1 socket.
Created attachment 1494034 [details] sockets values swapped with cores values
It looks like you are seeing the code working as intended now. OpenStack flavors don't model multi-core CPUs. The bug reported was that we were failing to show cores at all, since the existing code that set the number of cores per socket to one for OpenStack flavors was missing in the newly redesigned refresh code. If a flavor has "4 CPUs", this gets pulled into ManageIQ as 4 CPUs with one core per CPU.
Verified in 5.10.0.28. Cores and sockets are displayed correctly.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2019:0212