Build fails with "getaddrinfo() thread failed to start"

We are using the muslrust image for our CI builds. These were recently upgraded from Ubuntu Xenial to Jammy. After this upgrade, CI is failing with a strange error getaddrinfo() thread failed to start, and later could not execute process rustc -vV (never executed). I only managed to reproduce the error on cloud_drone_io and my selfhosted drone_yerbamate_ml, but not locally. Now the latter seems to have fixed itself after some OS upgrades + reboot.

Here is the issue in muslrust repo

Do you have any idea how this problem could be fixed?

Hi,

we are seeing the same issues. I suspect this might be a syscall incompatibility between the drone host and the guest libc. Cloud runs 4.15.0-50-generic. Were you able to solve this on cloud.drone ?

Cloud infrastructure is basically broken when using a newer distro like Ubuntu Jammy. Even had this error while running as root:

rm -rf ./bin
make: rm: Operation not permitted

Running the same pipelines but basing the image on focal runs fine

@jimsheldon any updates on this? cloud.drone.io is using a broken docker version, see this issue: Build and fetch errors with clux/muslrust:1.60.0 · Issue #94 · clux/muslrust · GitHub

@dessalines let me check in a while and get back to you.

Thanks!

@dessalines so this was the issue with the new ubuntu distribution that was updated without upgrading docker.

Hence, it is already sorted over the link you have shared. This has been labelled invalid and closed.

Anything else that I can help you with?

Thanks!

Just tested this, and the cloud drone is still using the broken docker version as of today: Drone CI

@dessalines let me get this checked and get back to you.

Thanks!

1 Like

What version of docker and docker-compose is cloud.drone currently running?

Any updates on this? Our builds have been broken for a while.

@dessalines I totally understand!

@brad can we get some help here?

Thanks

1 Like

Apologies for the delay here, we are having trouble tracking down someone who has access to these instances.

However, I think it is a safe bet that the version of docker is older, and that is causing the problem you are seeing.

It looks like you brought up your own drone instance at drone.join-lemmy.org? Do you expect to keep that instance going forward, or would you rather move back to cloud.drone.io?

We’d def prefer to move back to cloud.drone.io , we had to spin that up occasionally due to some issues with arm builds in the past.