-
Notifications
You must be signed in to change notification settings - Fork 2.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Increased latency when using bridge network with v5 #22786
Comments
I assume you run rootless. I would guess the delay is the dns resolution? Can you check by connecting directly to a ip? If it is dns please check /etc/resolv.conf on your host and in the container and provide the output.
This version is old and there have been several dns fixes in the meantime so I suggest you try this with a nwer pasta (passt) version installed, i.e. from https://copr.fedorainfracloud.org/coprs/sbrivio/passt/ |
(Warning: you may assume I know pretty much nothing in network "stuff" 😅 )
Right! Using IP, there's no delay.
In the container:
In the host:
(Obviously, the IP of the nameservers are internals to our network.)
Currently, we have Using Thanks! 👏 One thing unclear to me though is why the previous version of podman (v4.9.4) is still working with Out of curiosity, do you have references to the pasta issues you're referring to? |
Because podman 5 switched the default from slirp4netns to pasta. https://blog.podman.io/2024/03/podman-5-0-breaking-changes-in-detail/
No there are so many different ones that I honestly cannot tell without searching for a long time. |
Great, thanks a lot @Luap99 for the explanations and solution! |
Right, yes, I just happened to rebase the CentOS Stream 9 package from Fedora yesterday: but it might need a bit before mirrors pick up the new build.
It might be this by the way: https://passt.top/passt/commit/?id=d989eae308c2ea2032fc91cc04fb02dffe4a4b63 |
Issue Description
I noticed extra latency following the upgrade of podman version from 4.9.4 to 5.0.2.
I have a situation where doing a cURL to a server inside my network usually takes a few milliseconds but with podman 5.0.2 it now takes a few seconds.
This happens only when the container is using a bridge network.
For the record, this happens in a CI context where CI runs jobs in containers and I don't necessarily have full control on the containers/networks created.
Steps to reproduce the issue
1/ Having a podman network:
(in my case this network is created by the CI, no control on it)
2/ Run a container without linking it to the network and run a cURL:
https://www.google.com
or an internal serverOutput:
3/ Run the same but with a container linked to the network:
Output:
💥 Notice how the time is now 2 seconds instead of a few milliseconds.
Comparison with podman v4.9.4:
With container NOT attached to the network:
With container attached to the network:
That is similar response times no matter the podman network being used or not.
Describe the results you received
See above
Describe the results you expected
See above
podman info output
Podman in a container
No
Privileged Or Rootless
None
Upstream Latest Release
No
Additional environment details
VM with CentOS Stream release 9
Additional information
No response
The text was updated successfully, but these errors were encountered: