This repository has been archived by the owner on May 12, 2021. It is now read-only.
Set the queue size for Multiqueue virtio-net as the number of vCPUs on the guest. #719
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This api will allow the config to be accessed by other subsystems such as network. Signed-off-by: Archana Shinde <[email protected]>
Instead of using a default queue size of 8 for macvtap fds, use the number of CPUs on the guest as the queue size. This is the recommended approach. This also shown better performance results. Fixes kata-containers#680 Signed-off-by: Archana Shinde <[email protected]>
amshinde
force-pushed
the
net-queue-size
branch
from
c5c9445 to
63785ca
Compare
|
PSS Measurement: Memory inside container: |
|
Build failed (third-party-check pipeline) integration testing with
|
bergwolf
reviewed
Sep 13, 2018
| @@ -226,7 +225,7 @@ func networkLogger() *logrus.Entry { | |||
| func (endpoint *VirtualEndpoint) Attach(h hypervisor) error { | |||
| networkLogger().WithField("endpoint-type", "virtual").Info("Attaching endpoint") | |||
|
|
|||
| if err := xconnectVMNetwork(&(endpoint.NetPair), true); err != nil { | |||
| if err := xconnectVMNetwork(&(endpoint.NetPair), true, h.hypervisorConfig().NumVCPUs); err != nil { | |||
There was a problem hiding this comment.
Do we care about hotplugged CPUs? Properly not since they might be unplugged later. Just want to know your opinions about it.
There was a problem hiding this comment.
Hey @bergwolf - under what use-cases would we be actively hotplugging/unplugging?
There was a problem hiding this comment.
@egernst Currently no. I don't know any use case that would require removing a container while having other containers still running in the same sandbox. It is just that our APIs and CLI interfaces allow such behavior. nvm. It's fine with me to just use the default CPU numbers for the queue size.
egernst
approved these changes
Sep 13, 2018
|
LGTM! |
egernst
pushed a commit
to egernst/runtime
that referenced
this pull request
Feb 9, 2021
This includes fix for CVE-2019-19921 3291d66b rootfs: do not permit /proc mounts to non-directories 55f8c254 temporarily disable CRIU tests 5c20ea14 fix merging kata-containers#2177 and kata-containers#2169 8541d9cf Fix race checking for process exit and waiting for exec fifo 52951a7c Fix race in tty integration test with slow startup 8ddd8920 libcontainer: add method to get cgroup config from cgroup Manager cd7c59d0 libcontainer: export createCgroupConfig ec49f98d fs2: support legacy device spec (to pass CI) 88e8350d cgroup2: split fs2 from fs 41a20b58 Expose network interfaces via runc events 48b055c4 Makefile: allow overriding `docker` command 42690e68 Make event types public faf1e44e cgroup2: ebpf: increase RLIM_MEMLOCK to avoid BPF_PROG_LOAD error ccd4436f .travis.yml: add Fedora 31 vagrant box (for cgroup2) faf673ee cgroup2: port over eBPF device controller from crun 74a3fe5d cgroup2: do not parse /proc/cgroups 9c81440f cgroup2: allow mounting /sys/fs/cgroup in UserNS without unsharing CgroupNS 13919f5d Remove the static_build build tag. dbd771e4 cgroup2: implement `runc ps` 9996cf7d README.md: clarify cgroup2 support is not ready for production d918e7f4 cpuset_v2: skip Apply when no limit is specified 033936ef io_v2.go: remove blkio v1 code a610a848 criu: Ensure other users cannot read c/r files b28f58f3 Set unified mountpoint in find mnt func f017e0f9 checkpoint: Set descriptors.json file mode to 0600 4be50fe3 SECURITY: Add Security Policy 2111613c VERSION: back to development 28e58a0f Support different field counts of cpuaact.stats e63b797f Handle ENODEV when accessing the freezer.state file 5e0e67d7 fix permission denied 056909bd Adds note about user ns for rootless containers Fixes kata-containers#719 Signed-off-by: Archana Shinde <[email protected]>
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Instead of using a default queue size of 8 for Multiqueue virtio-net,
use the number of vCPUs on the guest as the queue size.
This is the recommended approach for setting up the queue size.
This also shown better performance results.