Red Hat Bugzilla – Bug 1026367
openshift-iptables-port-proxy service script should enable net.ipv4.conf.all.route_localnet
Last modified: 2017-01-25 01:38:55 EST
Description of problem:
The openshift-iptables-port-proxy service script will not function properly if net.ipv4.conf.all.route_localnet is disabled. Typically this is a step performed as part of an OpenShift deployment. However, as a safety check it should be temporarily enabled in this service script.
I say temporary because admins in different environments may have their own procedure for persisting sysctl changes. This change will warn the admin to persist the setting and mention that it is being temporarily enabled.
Commits pushed to master at https://github.com/openshift/origin-server
Bug 1026367 - openshift-iptables-port-proxy service script should enable net.ipv4.conf.all.route_localnet
Bug 1026367 - starting openshift-iptables-port-proxy after the network is up
Bug 1026367 - Improving oo-admin-ctl-iptables-port-proxy "UI"
Previously the restart, reload, stop and status commands were noop. This can
be confusing for admins.
stop: flushes the rhc-app-comm chain
restart: calls stop then start
reload: calls start
The motivation for this is the following:
1 Admin sees something weird with the rhc-app-comm chain not working as expected
2 Admin runs `for s in 'iptables network openshift-iptables-port-proxy'; do server $s restart'
3 Admin notices rhc-app-comm is still empty and is really confused
In this case the event that trigged #1 is that net.ipv4.conf.all.route_localnet
wasn't set in /etc/sysctl.conf.
Bug 1026367 - Moving the sysctl logic to oo-admin-ctl-iptables-port-proxy