d1nuc0m | tkajinam, today I was able to test the suggestion about nova pci setup - the compute node keeps on working, while the controller node does not | 13:06 |
---|---|---|
tkajinam | hmm that's wired | 13:07 |
tkajinam | d1nuc0m, so the current problem is that [pci] alias is not rendered in controller nodes , right ? | 13:13 |
tkajinam | for controller nodes you have to update your manifest to make sure that nova::pci class (or nova::api which includes nova::pci) is included seeing what you wrote in https://paste.opendev.org/show/bpjzI74Wm4pCrUlzMC5O/ | 13:15 |
tkajinam | I might be saying something stupid, though, as I might forget the important context behind that (sorry but it has been some time since we discussed this last time) | 13:16 |
d1nuc0m | Don't worry, I'm sorry for the delay, had the testbed down (again, IK) | 13:16 |
d1nuc0m | tkajinam, I am pretty sure classes are being included, as report_in_placement is correctly set to false, and this should only happen if I'm including nova::compute::pci in the same clause | 13:20 |
d1nuc0m | See here https://paste.opendev.org/show/bP8L0IxKHHGfSOC5hBIj/ | 13:20 |
d1nuc0m | Also tried to remove the conditional at all, with same result https://paste.opendev.org/show/bWjf6ChnvaUeuTAiKsiu/ | 13:23 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!