Feature #3228
closed
Please add an "Interface Down" Trigger Level for Gateway Group members
Added by Malte Stretz about 11 years ago.
Updated about 10 years ago.
Description
For gateways which are tied to interfaces (which are in all the cases I use Interface Groups) it would be nice to have a new trigger Interface Down which switches only if the actual underlying interface went offline/disconnected. This would work nicely with unpingable gateways without having to do hacks with extra IPs and fixed routes. (I actually always thought the Member Down was actually an Interface Down since the difference to Package Loss wasn't clear to me, now I found the explanation here https://doc.pfsense.org/index.php/Multi-WAN_2.0#Trigger_Level).
I don't think this would work with apinger; so it would need another trigger. I just had a quick look but maybe looking /tmp/${if}up in /etc/inc/gwlb.php is already enough.
Thanks!
- Status changed from New to Rejected
Interface link down will trigger "interface down" from monitoring, so I fail to see a point.
Chris Buechler wrote:
Interface link down will trigger "interface down" from monitoring, so I fail to see a point.
Not sure what you mean with monitoring, is that something which wasn't there or integrated with Multi-WANs in 2.0?
I must admit that I am also not sure what exactly I meant back in 2013. But I think the issue was with fail-over on a Multi-WAN setup with two PPoEs which didn't switch over to the secondary interface because the primary gateway was unpingable (due to the ISP blocking ICMP to the gateway) and thus an interface/gateway down wasn't detectable properly. The alternative is to set up some monitor IP somewhere on the internet, fix the route to that IP to one of the gateways and ping that. Unfortunately that's (a) a special firewall/routing rule just for some gateway up check and (b) then your interface would switch over if for some reason that host happened to go down.
IIRC in essence what I meant here was that it would be nice if you could configure a Multi-WAN setup to switch over when the underlying interface went down instead of requiring any ping magic at all.
Also available in: Atom
PDF
Updated by 
Updated by