-
Notifications
You must be signed in to change notification settings - Fork 2.4k
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
podman run returns error "unexpected end of JSON input" #11837
Comments
+1, experiencing the same issue on fresh install:
in my case the issue is on MacOs Big Sur. Downgrading to 3.3.1 helped. |
It's a bug in |
This comment has been minimized.
This comment has been minimized.
Duplicate of #11413 |
@Luap99 I tried to run with Error: error preparing container 89fed3a2d595dc28c4c5ce1af577b438a19457bcf0ddd9e1b19f9e658074d7aa for attach: error configuring network namespace for container 89fed3a2d595dc28c4c5ce1af577b438a19457bcf0ddd9e1b19f9e658074d7aa: error adding pod jovial_raman_jovial_raman to CNI network "podman": unexpected end of JSON input |
|
I suppose there is no way of getting a fixed podman-plugins rpm out ? 3.3.1-3
https://src.fedoraproject.org/rpms/podman/c/6972e15200c72420008cb71f4f8d1a483ee2aa9e?branch=rawhide containers/podman-machine-cni@0749884
So that one can patch the CoreOS installation so that it works again... (CNI) |
@Luap99 any idea when podman 3.4 will land in CoreOS? Current experiencing this issue on a fresh podman install |
Fedora CoreOS updates every two weeks (14 days): https://getfedora.org/en/coreos?stream=testing - see #11680 |
It looks like 3.4 is still working its way into both Fedora 34 and 35, which I assume needs to happen before FCOS picks it up: |
podman 3.4 landed in coreos next-devel a few hours ago, so it should be in the upcoming build for the next stream: https://builds.coreos.fedoraproject.org/browser?stream=next-devel&arch=x86_64 |
Any update on when this fix will be available for download from homebrew? |
I don't think |
Do we need to |
It will update every two weeks, you can check in the os-release (with ssh) podman 3.3.1 podman 3.4.0 This is supposed to be documented somewhere, and link to Fedora CoreOS And yes, the older podman release picks the newer operating system release... It seems like it is not possible to select the "next-devel" nor "testing-devel" streams. 😔 But maybe this could be a workaround ? https://getfedora.org/en/coreos?stream=stable EDIT: Sadly, no. It has the same broken |
That still uses a build from 2021-09-19:
Using
|
However, |
Where "newer" is supposed to be "better" ? (In this case, "older" would have been preferred). I was playing with using Fedora Cloud for podman instead of Fedora CoreOS, while it is broken. See https://github.com/afbjorklund/fedora-lima Same QEMU, Same RPMS, different distribution |
A manual image path can be provided: Taking the latest testing-devel qemu build from: https://builds.coreos.fedoraproject.org/browser?stream=testing-devel&arch=x86_64
From there I can run
|
That seems like the best workaround so far, should probably be featured on the home page |
Additionally, I can confirm that running the testing-devel version of coreos with podman 3.4 via the temporary workaround above resolves the "unexpected end of JSON input" errors I've been encountering! |
just a +1 that I encountered the same issue "error configuring network namespace for container ..." and the workaround fixed it |
If you're on M1 |
podman 3.4.0 is now available in the coreOS image. The workarounds mentioned above are no longer required. |
As Podman defaults to [1] https://docs.podman.io/en/latest/markdown/podman-machine-init.1.html#image-path |
Maybe the "image path" (i.e. CoreOS stream) should show in the machine list ? (especially since it changed from 3.3 to 3.4) |
Please open an issue to show this. |
Reported as: |
I am 100% in sympathy that this is a tricky problem, but I want to point out that these kinds of issues make it very hard to convince developers to use podman instead of docker. For all docker's faults, they do the exact same thing on Mac (run a VM machine) but they never asked the user to determine which machine image to use. If a new version of podman requires an updated machine image, then it should detect the problem and if not fix it, provide instructions to the user much better than this error message. |
You are describing close source versus open source. And you may be right, but I think Podman and Fedora CoreOS will improve and still allow flexibility for users to us other distributions. At one time Solaris worked better the Linux as well... |
I think some (including myself) was looking for a "slower" track, but it seems like that this would be covered by using Red Hat Enterprise Linux rather than Fedora ? The machine is handled by CoreOS and not by Podman, so I guess that's where to fix. i.e. Podman doesn't link to a particular version of the server, but uses the latest tested and available from the other project. Which means that any bug fix is first landed upstream, packaged in Fedora (RPM) and then included next fortnight (CoreOS) Any user wanting to get bugs fixed faster than this monthly cadence, would probably have to run their own VM instead ? Possibly there could be access to pre-releases or nightly builds, but eventually someone needs to support it as a "product" There is talk about providing a Podman Desktop, so better to address those concerns there instead of in this bug report: One way to get fixes faster is to run Fedora Cloud in the VM instead of Fedora CoreOS, as shown in various guides. Then you can administer and update the podman server yourself, through the package manager or even by building it ? |
Is this a BUG REPORT or FEATURE REQUEST? (leave only one on its own line)
/kind bug
Description
When I try to run a container using
podman v3.4.0
on macOS 11.6 x86_64, I get the following issue:Running a container fails for
podman remote
as well as when I ssh into the machine and execute the command with the same error.Describe the results you received:
Running a container fails with the error message:
Describe the results you expected:
The container starts.
Additional information you deem important (e.g. issue happens only occasionally):
Looking at the symptom this looks similar to #11452 and #11235, but I already checked and I do not have a
~/.docker
directory anymore.Result
Output of
podman version
:Output of
podman info --debug
:Package info (e.g. output of
rpm -q podman
orapt list podman
):Have you tested with the latest version of Podman and have you checked the Podman Troubleshooting Guide? (https://github.com/containers/podman/blob/master/troubleshooting.md)
Yes
Additional environment details (AWS, VirtualBox, physical, etc.):
The text was updated successfully, but these errors were encountered: