- Feature Articles
- CodeSOD
- Error'd
- Forums
-
Other Articles
- Random Article
- Other Series
- Alex's Soapbox
- Announcements
- Best of…
- Best of Email
- Best of the Sidebar
- Bring Your Own Code
- Coded Smorgasbord
- Mandatory Fun Day
- Off Topic
- Representative Line
- News Roundup
- Editor's Soapbox
- Software on the Rocks
- Souvenir Potpourri
- Sponsor Post
- Tales from the Interview
- The Daily WTF: Live
- Virtudyne
Admin
Cheers!
Admin
It took them 3 months to fix a peering issue that caused all of our data traffic to be routed across the country and back. I had numerous conversations with various "technicians" who were mostly clueless as to what was going on. I provided them with trace routes showing that packets were being routed from San Antonio (where our office was) to Dallas, then Chicago, then Washington D.C., then Atlanta, then back to Dallas (where some of our servers were located), but they insisted that it wasn't a problem with their routing, but a problem with our equipment. Even sending data to "close" servers a few miles away, or between my apartment and the office (less than 5 miles away), resulted in routes through DC, Chicago, or LA.
I tried to explain to the tech on the phone that they had a serious routing/peering issue and inquired as to whether or not a router at a peering point was down. Surely, I said, they peered with Level 3 (the bandwidth provider we used in Dallas) in Dallas? No, I was told, they did not peer with Level 3 in Dallas, but did peer with them in Chicago, LA, DC, and Houston.
But that wasn't the problem, according to the fine folks at XO. First they decided that the 130ms latency was because of a bad router, not the traversal of packets across the fucking country. Then after doing extended testing on the router, they decided that the problem was a wiring issue at the D-Mark. They sent a technician out and he couldn't find anything wrong with it. Then they decided that the problem was my router (which I had behind the Cisco 2600 they provided me so that I could do proper traffic shaping on the limited 1.5mbps so that I could actually keep good latency on the T1, assuming I could ever get it).
Finally, a tech told me that what I needed to do was set up BGP on MY router so that I could force certain routes to be taken. I administer the infrastructure for a small office of 15 people; I don't run a fucking datacenter. I should not have to set up BGP to get sub-100ms latencies to reasonably close machines.
I threatened to cancel our T1 about 3 times during the following 3 months. Eventually, the problem went away and we suddenly had 17ms latencies to Dallas and trace routes that showed our traffic hitting only hops which were geographically "on the way" there.
XO never explained to me what they fixed. They are a crappy company with a crappy infrastructure and terrible customer service. Fuck XO.
Admin
Network down.. Listen Failed
Admin
this reminds me of a story: a company was having daily outages. they finally found the problem: the room with the router had no outlets, so someone had made a whole in the wall into the WASHROOM next door...and once a day, someone unplugged the router to plug in a HAIR DRYER!