Hmm, indeed, the container will be always Up:
# cat /usr/local/bin/origin-node-run.sh
#!/bin/sh
set -eu
conf=${CONFIG_FILE:-/etc/origin/node/node-config.yaml}
opts=${OPTIONS:---loglevel=2}
function quit {
pkill -g 0 openshift
exit 0
}
trap quit SIGTERM
if [ ! -f ${HOST_ETC}/systemd/system/docker.service.d/docker-sdn-ovs.conf
]; then
mkdir -p ${HOST_ETC}/systemd/system/docker.service.d
cp /usr/lib/systemd/system/docker.service.d/docker-sdn-ovs.conf
${HOST_ETC}/systemd/system/docker.service.d
fi
/usr/bin/openshift start node "--config=${conf}" "${opts}" &
while true; do sleep 5; done
The while loop at the end is endless, even if openshift exited. Strange
idea.
What's the point of this?
Apparently, just handle the SIGTERM signal, and send it to all processes in
the same group:
# ps x -o "%p %r %c" | grep 9003
9003 9003 origin-node-run
9012 9003 openshift
32347 9003 sleep
Maybe it lacks a test in the while loop to ensure at least "openshift" is
running?
_______________________________________________
users mailing list
users@lists.openshift.redhat.com
http://lists.openshift.redhat.com/openshiftmm/listinfo/users
