< ^ txt
Wed May 13 09:32:54 EDT 2015
Went to bed a little late last night, but slept OK. Weather is cool and sunny.
Goals:
Work:
- Continue to hold down the fort until Scott gets back tomorrow
Done.
- Work on vm backups
A little.
- Play with Python web dev stuff, if I get a chance
Didn't get a chance.
Home:
- Make Mouse 1 on FvwmIconMan do RaiseAndFocus
Done.
Questions:
- When pinging a down host, generally `ping` will issue no out for lost packets. However, I notice that in some circumstances (power off or network cable unplugged?), a different remote host (some router in our ISP's environment?) instead replies with "time to live exceeded". Why?
I expected the answer to be easy to find, but I'm starting to think either I'm not googling the right terms, or it's something particular to our/our ISP's environment.
Possibly the link state between the external interface of our router and its gateway (the internal interface of our ISP's router) dynamically creates some route. If we unplug the network cable between the two, or power off our router, or otherwise cause the link state on the ISP router's internal interface to change, we fall through to some route that exceeds the TTL for pings (or is an infinite loop inside our ISP). The route may be dynamic based on the link state.
Interestingly, the host from our ISP that returns the TTL exceeded message in the event of the link being down is not a box that's normally visible in a traceroute when the link is up.
But I'm speculating at this point, and don't know enough about routing protocols to estimate how close I am to the mark. A traceroute with an unplugged cable might clarify things....
Yes. I'm not sure what exactly is happening inside the ISP's router, but when I unplug the network cable from our device, while watching with mtr, I can see an infinite routing loop between the ISP router and some other, new IP on the ISP's network. Diagnostically useful for us, but I'm guessing it's a misconfiguration on the part of our ISP.
< ^ txt