Slashdot Mirror
Comcast's Xfinity Internet Service Is Down Across the US [Update] (theverge.com)
Readers share a report: Comcast's internet service, Xfinity, appears to be suffering an outage across the country. DownDetector.com shows it being down around the United States, including in large cities like San Francisco, Los Angeles, and Boston. So far, online reports don't suggest that TV service or home phones are affected. On Twitter, Comcast confirmed the outage. Adding, "Some customers are having issues with their XFINITY Internet service. We apologize & appreciate your patience while we work to fix." The company tweeted moments ago, "Our teams continue to monitor an external network issue. We apologize for the inconvenience -- will provide updates as we learn more." In another tweet, Comcast said the issue is nationwide.
Update: At 20:39 GMT on Monday, Comcast said it had resolved the issue.
Update: At 20:39 GMT on Monday, Comcast said it had resolved the issue.
It could be an outbound issue, like a DNS server. That would also explain why the outage covers multiple regions.
I am not your blowing wind, I am the lightning.
There's a big global DDOS attack going on right now.
Our whole shop is down and can't access Azure or TFS
http://www.digitalattackmap.com/#anim=1&color=0&country=ALL&list=0&time=17475&view=map
This looks like it was actually a Level3 outage. Level3 carries a lot of the nationwide internet traffic (including traffic to/from Comcast), and the outage there caused various ripple effects (you can't get there from here, or can only get there very slowly across this (now) overloaded path).
Very slow in Seattle from about 10:00 AM to noon. Back to normal now.
I call shenanigans. I have it on Good Slashdot Authority that people in Seattle only have access to dialup.
#DeleteChrome
If their DNS isn't working, telling them to use "OpenDNS" won't be of much help.
https://146.112.62.105/
#DeleteFacebook
What the hell is a "carrier lost" joke supposed t{#`%${%&`+'${`%&NO CARRIER
#DeleteFacebook
Damn fools! Why can't they learn! Never make anything rouge! Use vert or bleu instead!
#DeleteFacebook
If this is a common issue, you might want to consider putting a caching DNS server locally on your LAN that does recursive lookups against Google's servers, but caches them locally so they're fast (after the initial, uncached lookup, of course).
It could be an outbound issue, like a DNS server. That would also explain why the outage covers multiple regions.
It was only certain sites. Sites that had their own CDN's like Facebook and Steam were fine. It may have been a Level 3 issue. There were two interesting things reported via twitter:
Comcast indicating "external network issue": https://twitter.com/comcastcar...
Level 3 indicating "our network experienced a disruption affecting some IP customers due to a configuration error": https://twitter.com/Level3NOC/...
Further evidence of outage, outage statistics from DownDetector:
http://downdetector.com/status...
http://downdetector.com/status...
We'll make great pets
No surprise, it appears to be human error in Border Gateway Protocol information sharing. Again.
Whether deliberate or unintentional, human-introduced errors in BGP routing (typo or espionage?) have happened before, and from reports I've seen, happened again this morning. This isn't something easily bypassed, like using numeric IP addresses instead of DNS lookups. It's fundamental to the resiliency of the Internet by design. Too bad such a fundamental part of the Internet architecture is still so dependent on trust.
In theory, practice and theory are the same. In practice, they rarely are.
Great catch here! It looks very much like someone at Level 3 screwed up the BGP tables introducing a much more specific /15 CIDR where a less specific /12 was already in place, which encapsulated a huge portion of Comcasts regional traffic and re-routed it through Level 3, which could not handle the load. Bonus points if anyone has the footage of the Level 3 tech getting flogged in the Noc ;)
/* * pope1 */