The trouble description seems a bit off. They had a backbone switch in San Jose fail (according to mtr), but it only affected DNS depending on your routing. In my case, 8.8.8.8 and 8.8.4.4 were both on the working side of the switch, but a lot of other sites and systems weren't -- even when accessing by IP. So it wasn't really a DNS issue.
I couldn't ping Hurricane Electric servers, but Slashdot worked OK for example.
The frustrating thing was that I was going through a Sacramento hop first but still couldn't reach systems in Reno -- I thought there was fiber all the way up the 80 corridor, but apparently most of my Comcast traffic has to go to San Jose before it can proceed East.
but a lot of other sites and systems weren't -- even when
accessing by IP. So it wasn't really a DNS issue.
That's just an intermittent issue that Comcast customers deal with--I experienced that in Indiana a couple of months ago.
I believe that it relates to them re-jiggering routers such that the appropriate default gateway for your modem changes, but because your modem is already up, it doesn't get notified of the change. So the modem doesn't send the message to the correct router, and the router it sends it to may or may not know how to get it to the destination.
I typically have logs streaming through terminal windows from several servers and multitask online -- I notice even brief transient network issues. FWIW what you're describing hasn't been a problem for me (though I don't doubt it is for lots of customers), and the San Jose outage was definitely out of the ordinary.
Anyway, I wish Sacramento was a bigger network hub, is all.
I couldn't ping Hurricane Electric servers, but Slashdot worked OK for example.
The frustrating thing was that I was going through a Sacramento hop first but still couldn't reach systems in Reno -- I thought there was fiber all the way up the 80 corridor, but apparently most of my Comcast traffic has to go to San Jose before it can proceed East.