CentOS 6 images instantly stop (Failed to open dir "/sys/fs/cgroup/blkio//lxc.payload/container-name")

After a recent host OS change (to Manjaro), CentOS 6 images, are no longer starting correctly. All other images, including CentOS 7, ubuntu, fedora, alpine etc appear to work fine.

LXD 3.16 installed with snap. brtfs storage pool, but also tested with a directory storage pool.

How could I go around troubleshooting this further? Log below:

# lxc launch images:centos/6
Creating the container
Container name is: relieved-phoenix
Starting relieved-phoenix

# lxc info relieved-phoenix --show-log
Name: relieved-phoenix
Location: none
Remote: unix://
Architecture: x86_64
Created: 2019/08/14 11:59 UTC
Status: Stopped
Type: persistent
Profiles: default

Log:

lxc relieved-phoenix 20190814115931.717 WARN     cgfsng - cgroups/cgfsng.c:chowmod:1525 - No such file or directory - Failed to chown(/sys/fs/cgroup/unified//lxc.payload/relieved-phoenix/memory.oom.group, 1000000000, 0)
lxc relieved-phoenix 20190814115933.404 ERROR    utils - utils.c:recursive_destroy:1782 - No such file or directory - Failed to open dir "/sys/fs/cgroup/blkio//lxc.payload/relieved-phoenix"
lxc relieved-phoenix 20190814115933.404 WARN     cgfsng - cgroups/cgfsng.c:cgroup_rmdir:1045 - Failed to destroy "/sys/fs/cgroup/blkio//lxc.payload/relieved-phoenix"

@jacob I’ve had problems with CentOS 6 containers on Alpine linux hosts. It might be related (or it may not), but I fixed it by adding this to the kernel boot parameters:

vsyscall=emulate

Also see

Thanks a lot @tomp, that seems to work perfectly!

1 Like