I’m running Silverblue/Rawhide and after a certain update rpm-ostree upgrade always gets stuck. It downloads several files (~10) and then the download stops and ostree takes >25% of CPU.
To update I need to go to the last working snapshot (Oct 28th), make the update there and switch to the newly installed snapshot.
I’ve got: ostree-2018.9-1.fc30.x86_64 and rpm-ostree-2018.9-3.fc30.x86_64
Any idea how to debug it? rpm-ostree doesn’t really give me a lot of hints what could be wrong.
Hmm. Can you file an issue against rpm-ostree for this? The forum here is OK for “broader” issues but this one is clearly a bug in the rpm-ostree stack. Can you try e.g. using rpm-ostree usroverlay, then download different ostree builds and use rpm -Uvh --force, and then systemctl restart rpm-ostreed and see if that helps?
Also, if you can reproduce this with plain ostree pull that would help. Since this issue is around pulling/networking it’s also likely reproducible in a container; rather than messing with the host you could use podman and create a container, then yum -y install ostree, ostree --repo=repo init, then ostree remote add ... etc.
One other useful command here is e.g. env OSTREE_DEBUG_HTTP=1 ostree pull fedora-silverblue:fedora/29/x86_64/silverblue