default network example does not work on RHEL

[KlavsKlavsen]

If I do:

    network_route { '10.0.0.0/8':
      ensure    => 'present',
      gateway   => '192.168.63.1',
      interface => 'eth0',
      netmask   => '8',
      network   => '10.0.0.0'
    }
    network_route { '192.168.0.0/16':
      ensure    => 'present',
      gateway   => '192.168.63.1',
      interface => 'eth0',
      netmask   => '16',
      network   => '192.168.0.0'
    }
    network_route { 'default':
      ensure    => 'present',
      gateway   => '192.168.63.50',
      interface => 'eth0',
      netmask   => '',
      network   => 'default'
    }

I get this in /etc/sysconfig/network-scripts/route-eth0:

10.0.0.0/255.0.0.0 via 192.168.63.1 dev eth0
192.168.0.0/255.255.0.0 via 192.168.63.1 dev eth0
default via 192.168.63.50 dev eth0

Which doesn't actually work :(

The only valid format for that file, that I know of, is:

GATEWAY1=192.168.63.1
ADDRESS1=192.168.0.0
NETMASK1=255.255.0.0

GATEWAY2=192.168.63.1
ADDRESS2=10.0.0.0
NETMASK2=255.0.0.0

and unlike what I'd normally expect (like what firewall module does) - the routes are NOT setup at puppet agent runtime - but one needs to reboot (or reload networking).

The resulting routes, after reboot, looks like this:

192.168.63.0    0.0.0.0         255.255.255.0   U     0      0        0 eth0
169.254.0.0     0.0.0.0         255.255.0.0     U     1002   0        0 eth0
192.168.0.0     192.168.63.1    255.255.0.0     UG    0      0        0 eth0
10.0.0.0        192.168.63.1    255.0.0.0       UG    0      0        0 eth0
0.0.0.0         192.168.63.1    0.0.0.0         UG    0      0        0 eth0

As far as I can see, the conclusions seem to be:

1. one cannot set default gw in routes-ethX (it's set in the ifcfg-ethX file)
2. ip commands like the ones used, does not work, and should probably use the officially supported format (or may it's just done wrongly in the module example or something - and is actually supported) ?
3. changes are not enacted on puppet run - one must reboot to update routes :(

[KlavsKlavsen]

I had to, manually, remove GATEWAY setting in /etc/sysconfig/network-scripts/ifcfg-ethX - and then the default network_route works as expected.
Setting the ip for eth0 interface with network_config - didn't even remove the GATEWAY setting.

Oddly enough, the other routes now works as well:

192.168.63.0    0.0.0.0         255.255.255.0   U     0      0        0 eth0
169.254.0.0     0.0.0.0         255.255.0.0     U     1002   0        0 eth0
192.168.0.0     192.168.63.1    255.255.0.0     UG    0      0        0 eth0
10.0.0.0        192.168.63.1    255.0.0.0       UG    0      0        0 eth0
0.0.0.0         192.168.63.50   0.0.0.0         UG    0      0        0 eth0

(I rebooted once more).

[adrienthebo]

Thanks for reporting this!

Looking at https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Deployment_Guide/s1-networkscripts-static-routes.html , it seems like the format for the routes should be valid; if I'm understanding you correctly then the routes did start working?

The documentation also says that you can add a default route of the format default via X.X.X.X dev interface; is the problem that GATEWAY was added to the file and then not removed later? Were you managing the interface file with the network_config type?

With respect to not applying the configuration immediately, I've left that up to users to handle. Restarting networking is something that you can't get wrong - if Puppet was restarting the network each time a route was updated it's entirely possible that the machine could get knocked offline. Since it's a bit risky, I've opted to not manage that and let end users do what best suits them. Does that make sense?

[gwarf]

Hi,
The format is not valid, when using IP commands arguments format, the netmask should be specified using CIDR notation:

10.8.0.0/24 via 172.20.8.1 dev eth1

Instead of

10.8.0.0/255.255.255.0 via 172.20.8.1 dev eth1

If not the network init script returns an error:

Bringing up interface eth1:  Error: an inet prefix is expected rather than "10.8.0.0/255.255.255.0".
Error: an inet prefix is expected rather than "10.8.1.0/255.255.255.0".

And the module is always removing the CIDR notation from the route, even when setting netmask to 24.

[gwarf]

Hi,
I managed to get this working using the changes mentionned in #58.

[confiq]

I can confirm it's working well...
It's been one year and half and nobody merge this

[adrienthebo]

@confiq if the pull request worked for you, could you update the pull request with tests and rebase it so that it can be merged?

[confiq]

@adrienthebo: are we talking about PR #62 ?
It's not mine PR so I can't rebase it but if needed I can make new PR.
About tests, It passes all tests so I don't see a problem merging it. Unless you want to write tests for new check new CIDR notations

CC @gwarf

[adrienthebo]

A new PR would be fine; I'm fine with merging code to fix this issue but the existing PR has merge conflicts. Since there is a behavior change, it does need tests before it gets merged.

[adrienthebo]

And if @gwarf is willing to continue working it then superb, but if that doesn't work out I'm happy to accept PRs from other locations.

[confiq]

Ok,
I've been using HEAD version and it's working fine.
I can't confirm this bug anymore.

[igalic]

okay, thank you @confiq!