We noticed cgroup doesn’t honor user namespace remap. This breaks systemd to
be running inside container.
With userns-remap set to default.
root@lnx-01 docker]# docker logs cdh_SR1111
systemd 219 running in system mode. (+PAM +AUDIT +SELINUX +IMA -APPARMOR
+SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 -SECCOMP
+BLKID +ELFUTILS +KMOD +IDN)
Detected virtualization kvm.
Detected architecture x86-64.
Welcome to Oracle Linux Server 7.6!
Set hostname to <cdh_SR1111.us.oracle.com>.
Failed to read AF_UNIX datagram queue length, ignoring: No such file or
directory
Failed to create root cgroup hierarchy: Permission denied
Failed to allocate manager object: Permission denied
[!!!] Failed to allocate manager object, freezing.
[root@lnx-01 docker]#
[root@lnx-01 docker]# docker exec -i cdh_SR1111 ls -l
/sys/fs/cgroup
total 0
dr-xr-xr-x. 6 65534 65534 0 Jan 17 06:16 blkio
lrwxrwxrwx. 1 65534 65534 11 Jan 17 04:50 cpu -> cpu,cpuacct
dr-xr-xr-x. 6 65534 65534 0 Jan 17 06:16 cpu,cpuacct
lrwxrwxrwx. 1 65534 65534 11 Jan 17 04:50 cpuacct -> cpu,cpuacct
dr-xr-xr-x. 4 65534 65534 0 Jan 17 05:25 cpuset
dr-xr-xr-x. 6 65534 65534 0 Jan 17 06:16 devices
dr-xr-xr-x. 4 65534 65534 0 Jan 17 05:25 freezer
dr-xr-xr-x. 4 65534 65534 0 Jan 17 05:25 hugetlb
dr-xr-xr-x. 6 65534 65534 0 Jan 17 06:16 memory
lrwxrwxrwx. 1 65534 65534 16 Jan 17 04:50 net_cls -> net_cls,net_prio
dr-xr-xr-x. 4 65534 65534 0 Jan 17 05:25 net_cls,net_prio
lrwxrwxrwx. 1 65534 65534 16 Jan 17 04:50 net_prio -> net_cls,net_prio
dr-xr-xr-x. 4 65534 65534 0 Jan 17 05:25 perf_event
dr-xr-xr-x. 6 65534 65534 0 Jan 17 06:16 pids
dr-xr-xr-x. 2 65534 65534 0 Jan 17 04:50 rdma
dr-xr-xr-x. 6 65534 65534 0 Jan 17 06:16 systemd
with userns-remap disabled
===================
root@ci-phx-cdhdevinfra-ad1-lnx-01 docker]# docker logs cdh_SR1111
systemd 219 running in system mode. (+PAM +AUDIT +SELINUX +IMA -APPARMOR
+SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 -SECCOMP
+BLKID +ELFUTILS +KMOD +IDN)
Detected virtualization kvm.
Detected architecture x86-64.
Welcome to Oracle Linux Server 7.6!
Set hostname to <cdh_SR1111.us.oracle.com>.
[ OK ] Reached target Paths.
[ OK ] Reached target Local File Systems.
[ OK ] Created slice Root Slice.
[ OK ] Listening on Delayed Shutdown Socket.
[ OK ] Listening on Journal Socket.
[ OK ] Reached target Swap.
[ OK ] Created slice System Slice.
Starting Create Volatile Files and Directories…
[ OK ] Reached target Slices.
[root@lnx-01 docker]#
Docker run command used.
docker run -itd --tmpfs /tmp --tmpfs /run --cap-add=SYS_ADMIN -v
/sys/fs/cgroup --cap-add=SYSLOG --restart=unless-stopped --cpus=2.0 -m 10.0GB
–cap-add=NET_RAW --cap-add=NET_ADMIN -p 21000:22 -p 31000:5901
–name=cdh_SR1111 -h cdh_SR1111.us.oracle.com --label CDH_INCIDENT=SR1111
–label CDH_CUSTOMER=bofa --label CDH_BUCKET=nsin08 -e CDH_INCIDENT=SR1111
–label CDH_LOGDIR=/saascdh/download/ -e CDH_UNIX_GROUP=CDHSR1111 --label
CDH_REGION=us-phoenix-1 test_docker:oel7-dev
This issue seems similar to one described in
https://bugzilla.redhat.com/show_bug.cgi?id=1406684 but i couldn’t find
suggested fix for OEL 7.7
Host OS is oel 7.7 oci oke vm silver image for saas, docker image os is OEL
7.6
[root@ln-01 log]# uname -r
4.14.35-1902.9.2.el7uek.x86_64
% docker -v
Docker version 19.03.1-ol, build ead9442
So far i have tried mounting separate tmpfs as /sys/fs/cgroup, fix as suggested in https://bugzilla.redhat.com/show_bug.cgi?id=1566680, configuring oci-systemd-hook but none seems to work. I am out of options here, what else i can try to make it work?