DHCP fails on WAN interface multiple instances

[erikig]

I’ve got a strange one here. This morning, I woke up to discover my home pfSense box (2.6.0-RELEASE) had lost the DHCP supplied IP on the WAN interface. It’s a dual WAN configuration, so I was able to switch over to the secondary. I though the problem was with my ISP, but when I got to the office I discovered two other pfSense boxes with the same issue on different configurations (two hardware and one VM) and different ISPs. In cases where the machines were dual WAN, only the primary WAN interface was affected.

All of them seemed to stop working around midnight UTC (1AM CET). I tried restoring one here, but ran into the double <sshdata> block in the backup configuration file bug. With that fixed, I did a factory reset and was able to restore the configuration from the backup.

Which brings me back to where I started. Still no DHCP on the primary WAN interface.

Anyone else seen this kind of behavior?

[eableson]

@erikig And I figured out what was happening. A confluence of events. On the one side, yes the ISP’s DHCP server went offline although their front-line support kept insisting nothing was wrong. (Yes, I know the fibre is up, that’s not the problem. ).

What caused all of the odd secondary behavior was that the syslog server crashed and as a result pfSense started generating huge amounts of logs (notably system.log) which filled up the disk which resulted in all sorts of things breaking like DHCP etc.

Clearing out the excess in /var/log and rebooting put things back on track. Other than of course the original source of the problem which the ISP finally acknowledged 4 hours later with a generic “there’s an incident impacting your line”