xen-1.1.1: Skip virtual, non-NUMA net/blk,tun/tap and vif OSDevs #5
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
JFYI: This PR is only for showing the change in the version 1.1.1 of the Xen backend.
It fixes that the issue that when running load tests were run on large NUMA systems, several dozens of
tap
andvif
devices may be created and those would "spam" the data captured by lstopo.As those are not relevant to capture the HW topology, ignore those OSDevs that are purely virtual and are not associated with an actual HW device, and thuse are not relevant to capturing the NUMA topology.
For example, this is the ASCII art output from a 4 socket, 8 NUMA node system that would show several dozens of vif devices and some tap devices when load tests run that are not relevant and should be omitted like shown here.
The output is still not perfect as the NUMA PCI topology is not reported by Xen on this machine (also by
xl info -n
so it's not a hwloc issue), but this already a lot better than without this improvement.SVG output:
ssh -t [email protected] lstopo --of svg >08-04.svg
ASCII art output: