OLSR dynamic gateway plugin

[rebel2234]

did the "olsrd_dyn_gw.so.0.3" plugin make it into the 1.3.12 (build 2731) release? I cannot get it to work for the life of me. Is there something I am missing?

LoadPlugin "olsrd_dyn_gw.so.0.3"
{
PlParam "Interval" "5"
PlParam "Ping" "192.168.11.1"
PlParam "Ping" "172.18.0.1"
}

[tony]

It's in there, but use olsrd_dyn_gw.so.0.4 with this release, and you should be all set.

Other OLSR plugins available include:
olsrd_arprefresh.so.0.1
olsrd_dot_draw.so.0.3
olsrd_httpinfo.so.0.1
olsrd_txtinfo.so.0.1

[rebel2234]

One more quick question while im at it. This plugin detects if there is a valid gateway to the internet where would I use this plugin? At the endpoints of my network or somewhere else?

[tony]

You would add this to the system(s) that have Internet connections.

[DrLove73]

One more quick question while im at it. This plugin detects if there is a valid gateway to the internet where would I use this plugin? At the endpoints of my network or somewhere else?

Can this work with policy routing gateways (like only dest port 80), or just regular gateways?

[tog]

The dyngw plugin announces 0.0.0.0/0 via Hna4 based on whether or not there is a response from what you told it to ping.

Policy routing is pretty much a static route type entry and will not change based on anything that OLSR does.

[tony]

Policy routing gateways are a layer above the standard routing table (and a higher priority). OLSR, and the other routing daemons only manipulate the standard routing table.

[DrLove73]

I was asking this because I use policy routing for http traffic. If "http" gateway (or towards it) goes down, I am stuck with no service until I manualy turn http marking rule off.

[tog]

Yes.

Only way I can think of to get around that in any automated fashion would be to have a monitoring script upload and activate a pre-made failover "no policy route" config for you via starutil if it detected such an event.

You'd have to be really diligent about keeping your failover/normal configs up to date so your automated monitoring script didn't upload an old config that left out some important static route or other config change.

[DrLove73]

I already concluded that, but I will have to carefully plan where to put monitoring script.
It would be nice to have some if then possibility in Star-Os it self tied to few ping watchdog instances. A LOT of trouble would be avoided in that case.