ovs-p4ctl built with GitHub Actions crash with a segfault due to AVX512 support being enabled when run on Virtualbox or other systems lacking AVX support #118
Description
I just spent a few hours looking into an issue which turns out to be caused by the fact the images we are building and pushing to GHCR do not work.
Pull the GHCR Ubuntu 20.04 image like this:
docker pull ghcr.io/ipdk-io/ipdk-ubuntu2004-x86_64:main
Next, setup your ipdk.env file as shown:
vagrant@ubuntu-focal:~/ipdk/build$ cat ~/.ipdk/ipdk.env
#
# IPDK CLI configuration file
# See for more information https://github.com/ipdk-io/ipdk/blob/main/build/IPDK_Container/README_DOCKER.md#CLI-configuration-settings
#
# /home/vagrant/ipdk/build/scripts contains the location of the 'sourcing' ipdk script when read.
#
BASE_IMG=ubuntu:20.04
IMAGE_NAME=ghcr.io/ipdk-io/ipdk-ubuntu2004-x86_64
DOCKERFILE=/home/vagrant/ipdk/build/scripts/../Dockerfile.ubuntu
TAG=main
vagrant@ubuntu-focal:~/ipdk/build$
Star the container:
$ ipdk start -d
Login and see how ovs-vswitchd is not running, and when you start it manually, you see a segfault:
vagrant@ubuntu-focal:~/ipdk/build$ docker exec -it ipdk bash
root@f111597b0a04:~# ps axuw
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.0 0.0 3976 3244 pts/0 Ss+ 20:47 0:00 /bin/bash /root/scripts/start.sh rundaemon
root 45 0.0 0.0 2520 584 pts/0 S+ 20:47 0:00 sleep infinity
root 95 0.0 0.0 5300 2708 ? Ss 20:47 0:00 ovsdb-server --remote=punix:/usr/local/var/run/openvswitch/db.sock --remote=db:Open_vSwitch,Open_vSwitch,ma
root 100 0.5 0.0 4240 3668 pts/1 Ss 20:50 0:00 bash
root 114 0.0 0.0 5896 2836 pts/1 R+ 20:50 0:00 ps axuw
root@f111597b0a04:~# /root/scripts/start.sh rundaemon
Start as long running process.
1024
1024
Killing OVS Processes If Already Running....
Killing ovsdb-server process....
Creating /var/run/openvswitch
Starting OvS DB server....
Staring OvS VSWITCHD Process....
/root/scripts/run_ovs.sh: line 45: 153 Illegal instruction (core dumped) GLOG_log_dir=/tmp/logs ovs-vswitchd --pidfile --detach --no-chdir --mlockall --log-file=/tmp/logs/ovs-vswitchd.log
^C
root@f111597b0a04:~#
Now, I have pulled down the IPDK code and built a container locally, and when I run that image, it works fine:
vagrant@ubuntu-focal:~/ipdk/build$ docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
ghcr.io/ipdk-io/ipdk-ubuntu2004-x86_64 sha-bb7a773 9141307b21db 10 minutes ago 3.23GB
ghcr.io/ipdk-io/ipdk-ubuntu2004-x86_64 main 840c958df8e8 30 hours ago 3.23GB
ubuntu 20.04 54c9d81cbb44 3 weeks ago 72.8MB
vagrant@ubuntu-focal:~/ipdk/build$ cat ~/.ipdk/ipdk.env
#
# IPDK CLI configuration file
# See for more information https://github.com/ipdk-io/ipdk/blob/main/build/IPDK_Container/README_DOCKER.md#CLI-configuration-settings
#
# /home/vagrant/ipdk/build/scripts contains the location of the 'sourcing' ipdk script when read.
#
BASE_IMG=ubuntu:20.04
IMAGE_NAME=ghcr.io/ipdk-io/ipdk-ubuntu2004-x86_64
DOCKERFILE=/home/vagrant/ipdk/build/scripts/../Dockerfile.ubuntu
vagrant@ubuntu-focal:~/ipdk/build$ ipdk start -d
Loaded /home/vagrant/ipdk/build/scripts/ipdk_default.env
Loaded /home/vagrant/.ipdk/ipdk.env
Can't find update-binfmts.
Using docker run!
b9aee9f0aa34e94785023274592a48b9b3bd1e5e2d2f7b3fce6c2a8435e766ba
vagrant@ubuntu-focal:~/ipdk/build$ docker exec -it ipdk bash
root@b9aee9f0aa34:~# ps axuw
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.1 0.0 3976 3096 pts/0 Ss+ 20:50 0:00 /bin/bash /root/scripts/start.sh rundaemon
root 41 0.0 0.0 5300 2676 ? Ss 20:50 0:00 ovsdb-server --remote=punix:/usr/local/var/run/openvswitch/db.sock --remote=db:Open_vSwitch,Open_vSwitch,ma
root 45 97.0 2.3 67827184 194716 ? SLsl 20:50 0:07 ovs-vswitchd --pidfile --detach --no-chdir --mlockall --log-file=/tmp/logs/ovs-vswitchd.log
root 63 0.0 0.0 2520 596 pts/0 S+ 20:50 0:00 sleep infinity
root 69 0.0 0.0 4240 3488 pts/1 Ss 20:50 0:00 bash
root 83 0.0 0.0 5896 2876 pts/1 R+ 20:50 0:00 ps axuw
root@b9aee9f0aa34:~#