Hi Group Sorry guys I think this was a config error. I had the same address in the primary test and secondary route which I suspect broke things.
Regards Michael Knill Sent from my iPhone so please excuse my brevity. > On 13 Nov 2020, at 9:07 am, Michael Knill <michael.kn...@ipcsolutions.com.au> > wrote: > > Hi Lonnie > > Yes a weird one. It is a static IP Address on eth0.100 and the cable was out > for a while yet it did not fail over until I manually shut the interface. > This was even after I reset the box. > I looked at the routing table and the default route remained pointing to the > Default Gateway. > I will do some more testing and let you know. > > Regards > Michael Knill > > On 13/11/20, 2:45 am, "Lonnie Abelbeck" <li...@lonnie.abelbeck.com> wrote: > > Hi Michael, > > If I understand correctly, your upstream internet worked for some > destinations, but had an outage for others, so one of your WAN Failover -> > "Target IPv4 Hosts" still worked, so it did not automatically failover. > > Unless this was a one-off fluke, you may want to consider "Target IPv4 > Hosts" closer to your critical infrastructure. > > The fact that the WAN is a tagged VLAN should not make any difference. > > Back to your question, what happens when the WAN (External Interface) link > goes down while failover is running. In general it should force a failover. > > But keep in mind, in order for the failover script to work properly, > static /32 routes are defined for each "Target IPv4 Hosts" over the *primary* > interface (or ppp0 for PPPoE). Without these static routes as soon as the > secondary link became active the "Target IPv4 Hosts" would be reachable and > failover would return back to primary ... maintaining the "Target IPv4 Hosts" > static /32 routes over the primary interface (or ppp0 for PPPoE) is very > important internally to the failover script. > > If something would happen to remove the primary interface and any routes > associated with it, behind the back of the failover script, things would not > work as expected temporarily, but the failover script will automatically > refresh these routes if it detects they are missing. > > So "pulling the cable" may have forced failover if you waited long enough. > > Anyway, defining "Target IPv4 Hosts" closer to your critical > infrastructure may be a solution. > > Lonnie > > > > >> On Nov 12, 2020, at 5:00 AM, Michael Knill >> <michael.kn...@ipcsolutions.com.au> wrote: >> >> Hi Group >> >> I got caught this evening. >> One of our providers requires a tagged sub interface to be used for their >> WAN connection e.g. eth0.100. >> The WAN connection was broken but you could still ping at least one of the >> test IP Address. Ok that's fine I will just disconnect eth0 and it will fail >> over. Wrong it didn't! >> I had to shut down the sub interface with ‘ifconfig eth0.100 down’ to make >> it actually be down so it failed over. >> Is there a better way to do this? Can I actually bring this interface down >> externally any way? >> >> Thanks guys. >> >> Regards >> Michael Knill >> _______________________________________________ >> Astlinux-users mailing list >> Astlinux-users@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/astlinux-users >> >> Donations to support AstLinux are graciously accepted via PayPal to >> pay...@krisk.org. > > > > _______________________________________________ > Astlinux-users mailing list > Astlinux-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/astlinux-users > > Donations to support AstLinux are graciously accepted via PayPal to > pay...@krisk.org. > _______________________________________________ Astlinux-users mailing list Astlinux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/astlinux-users Donations to support AstLinux are graciously accepted via PayPal to pay...@krisk.org.
