0

We've got an Ubuntu 16.04.3 LTS machine plugged into our network that occasionally needs to automatically reboot itself. About 25% of the time, when it does this, it takes down the entire network. Both ethernet and wifi connections stop working.

Simply turning the machine off or unplugging its ethernet cord resolves the issue for all the other hosts.

The network is configured for DHCP and is powered by pfSense 2.3.5. The machine in question runs through two different switches (one near the router and a small one near the machine to expand the number of available ports).

The network is configured for DHCP, though many of the devices have "static leases" which is to say they are given the same IP every time they ask for an address; the machine in question is configured this way.

I can turn the machine back on and ssh in without issue thereafter.

How do I diagnose this issue? I'm not sure what I should be looking for. It'd be amazing if I could make my network resilient to this sort of denial-of-service in the future.

dave mankoff
  • 121
  • 5
  • Why is it connected to two switches? What services are on it? Telll us more about the system. – Zoredache Dec 05 '17 at 16:21
  • Added version info and a little more color. – dave mankoff Dec 05 '17 at 16:40
  • Not much info to go on. There are various ways to troubleshoot this: does DHCP work? can you ping the gateway? Does layer 2 work, e.g. look at the arp table, do you see a mac address for the gateway's IP address? Or try an arping. You can also try a wireshark or tcpdump to dump traffic when it goes down. – Mike Marseglia Dec 05 '17 at 20:23
  • Nothing seems to work. The only things I haven't tried is giving a machine a static IP, and plugging a monitor/serial console into the router directly during an outage. I'm hoping that there's something in the logs that I can suss out after I turn the offending machine off, but I'm not sure what to look for. – dave mankoff Dec 05 '17 at 21:50

0 Answers0