My use case is I'm running rootless podman, with systemd containers, and the btrfs storage drivers. Everything works fine on systemd 249.4-r4, after the new systemd 249.6 update podman start or run fails with the following types of errors: Error: OCI runtime error: unable to start container "b7dd10cb2380d744a25cd607fd34a47ec76924d178ebf1316b3bc5a7a44392c7": container_linux.go:380: starting container process caused: process_linux.go:385: applying cgroup configuration for process caused: error while starting unit "libpod-b7dd10cb2380d744a25cd607fd34a47ec76924d178ebf1316b3bc5a7a44392c7.scope" with properties [{Name:Description Value:"libcontainer container b7dd10cb2380d744a25cd607fd34a47ec76924d178ebf1316b3bc5a7a44392c7"} {Name:Slice Value:"user.slice"} {Name:PIDs Value:@au [63946]} {Name:Delegate Value:true} {Name:MemoryAccounting Value:true} {Name:CPUAccounting Value:true} {Name:IOAccounting Value:true} {Name:TasksAccounting Value:true} {Name:DefaultDependencies Value:false}]: error creating systemd unit `libpod-b7dd10cb2380d744a25cd607fd34a47ec76924d178ebf1316b3bc5a7a44392c7.scope`: got `failed`
I've upgraded systemd from 249.4-r4 to 249.6, and podman still works for me. Did you by chance disable the cgroup-hybrid USE flag? Note that you can switch back to cgroup v1 by adding systemd.unified_cgroup_hierarchy=0 to your kernel parameters.
Zac, now that you asked, I tried it with unified-cgroups emerged on and off as well as rebuilding initramfs (dracut via gentoo-kernel --config) and never did get it working on the latest systemd. In the end I ended up switching to LXD as it was better for my use case. Sorry, that definitely wasn't a solution but did fit my use case better. Thanks again for the help and hope that if this bug crops up again yall will have something to referrence.