quick favor

[belovedcrown.livejournal.com]

can someone here please run a tracert to one of my webservers from your location: family.parjen.com and email me the results?

send to jenniparks AT parjen.com

thanks. stupid #@$#@ comcast and their latency. $160 a month for a business connection and more latency then when i was on cox cable on a cheaper plan. overloaded nodes i tell you!!!!

Comments

[zig-mover.livejournal.com]

sent.

[belovedcrown.livejournal.com]

good! i am at your mercy then. are you tech level one can't make things happen or high up the chain power whom i should throw myself at and beg for mercy?

this is so freaking aggravating. 1 year with cox and NEVER a problem. move to new area where comcast is provider have to pay $160/mnth to get the same service that i got with cox for only $79. supposed to be 7mbps downstream and 712kbps upstream. instead i am getting dialup speeds and my websites are barely reachable from the outside...

see:

http://family.parjen.com/lj/jennihair2.jpg

[said-by-me.livejournal.com]

Comcast has been seriously sucking here too (Sacramento) Its gone down 3 times in the last 5 days. At one point being down for almost 24 hours. All I can get from them is that they are having issues with the node I am on. I went to comcast because everyone said how great they were. In 6 months I've had downtimes at least 6 times.

[gaymafiakingpin.livejournal.com]

What sort of troubleshooting have you done. I used to work for Comcast.

Power-cycled modem and router?
Rebooted?
Do you have trouble browsing or just with accessing your site?

If the nodes were so over-loaded that you were getting dial-up speeds Comcast would be getting a huge influx of calls and you would have waited on hold for days to talk to someone.

[abstrak-tokatl.livejournal.com]

no.

[agentz.livejournal.com]

Good grief, it took my machine just about 43 seconds to load that image, which is what, about 29kb? I'm on a 1.5Mb cable service in the UK.

[teriwyn.livejournal.com]

I -also- used to work for Comcast. :P Boyfriend still does.

The image loaded super-snappy for me, but I tracerouted and observed that the traffic didn't hit "the real internet" at all. Using Comcast in MD, all hops stayed on comcast IPs, all the way to your server. So, my traceroute is probably 100-million times useless to you.

BTW, the OP is on a comcastbusiness.net IP. comcastbusiness.net is just a bit different from what you and I are probably familiar with. I don't think she even calls the regular lines.

If this were a node or an RF issue, I wouldn't have been able to get such a snappy response time, since those bottlenecks would've appeared to anyone and everyone who tried to access it. So if others are getting el-crapo speeds, it's not something that will be found anywhere between the server and the UBR (Or cadent - I think they ditched UBRs in favor of Cadent routers in most areas.)

This screams 'bandwidth bottleneck' to me. I'll do some more poking about to see what else I can find.

[teriwyn.livejournal.com]

Er, wow. Try to make sense of my above post, even though I changed who I was directly speaking to about five times with no real clarification. A stupid is me.

[belovedcrown.livejournal.com]

thank you.

i've apparently isolated the problem but do not know what is causing it. At first the reps at comcast biz thought i was crazy because, following their script, they could ping my router and even my webservers just fine. i could do the same to them and tracert times were not to bad. However, when you up the packet size, you IMMEDIATELY saw the problem. i tried doing a ping -l 200 and they did the same and suddenly everyone is getting 75% packet loss. the problems seems to start not at my router or the first hop, but then next hop, a comcast router in alexandria. argh. they are "looking into" the problem. meanwhile i am virtually unable to get out such as through term serv in the middle of the night, the exact reason why my employer [us house of representatives] is paying the bucks for me to have a business grade connection with a service level agreement. this sucks all the way around and i miss cox cable internet.

[teriwyn.livejournal.com]

Yeah, the issue must be someplace in your path to the outside world. Since when I trace to you, all the packets stay inside Comcast's sandbox, I couldn't duplicate what you were referring to, even with -l 1000. Response time remained snappy and packets weren't dropped.

What's the IP of the troubling hop? I wonder if I can duplicate the issue by pinging it directly?

[teriwyn.livejournal.com]

http://www.dnsstuff.com/tools/tracert.ch?ip=family.parjen.com

There's a trace from the outside, but no way to vary the packet size, so I couldn't reproduce the problem.

Does that path reflect one of the problem hops?

[belovedcrown.livejournal.com]

neither of us will be able to detect the problem now, as its not occurring. :) as i mentioned in an earlier post, the problem comes and goes, usually occuring most frequently between 11pm and 8am.

thanks for all your help though. and this is the router where the packet loss seems to begin, a comcast router in reston. its the second hop for me:

C:\Documents and Settings\jparks>ping -l 200 68.87.129.114



Pinging 68.87.129.114 with 200 bytes of data:



Request timed out.

Reply from 68.87.129.114: bytes=200 time=32ms TTL=253

Reply from 68.87.129.114: bytes=200 time=17ms TTL=253

Request timed out.



Ping statistics for 68.87.129.114:

Packets: Sent = 4, Received = 2, Lost = 2 (50% loss),

Approximate round trip times in milli-seconds:

Minimum = 17ms, Maximum = 32ms, Average = 24ms



here is a full default packet size tracert for example showing my hops to comcast.net. just so you have an idea of the routers 1-2 hops away from me:

1 * * * Request timed out.

2 14 ms 13 ms 14 ms ge-1-1-ur01.manassas.va.bad.comcast.net [68.87.135.209]

3 14 ms 14 ms 14 ms te-9-2-ur01.reston.va.bad.comcast.net [68.87.129.114]

4 32 ms 15 ms 16 ms te-9-3-ur02.arlington.va.bad.comcast.net [68.87.128.161]

5 16 ms 16 ms 14 ms te-9-1-ur01.arlington.va.bad.comcast.net [68.87.128.157]

6 30 ms 16 ms 16 ms te-9-3-ur02.alexandria.va.bad.comcast.net [68.87.128.141]

7 32 ms 16 ms 14 ms te-9-1-ur01.alexandria.va.bad.comcast.net [68.87.128.133]

8 277 ms 298 ms 299 ms te-8-1-ur01.capitolhghts.md.bad.comcast.net [68.87.128.138]

9 16 ms 17 ms 16 ms te-8-1-ur02.capitolhghts.md.bad.comcast.net [68.87.129.150]

10 15 ms 18 ms 14 ms te-9-2-ar01.capitolhghts.md.bad.comcast.net [68.87.129.145]

11 16 ms 17 ms 16 ms 68.87.16.153

12 18 ms 17 ms 16 ms 12.118.122.9

13 81 ms 80 ms 91 ms tbr1-p010401.wswdc.ip.att.net [12.123.9.106]

14 80 ms 76 ms 76 ms tbr1-cl4.sl9mo.ip.att.net [12.122.10.30]

15 76 ms 75 ms 76 ms tbr1-cl2.sffca.ip.att.net [12.122.10.42]

16 78 ms 76 ms 115 ms 12.123.13.178

17 78 ms 78 ms 76 ms idf22-gsr12-1-pos-6-0.rwc1.attens.net [12.122.255.222]

18 79 ms 78 ms 77 ms rwcsbix12-3-1.attbi.com [63.241.85.242]

19 82 ms 141 ms 79 ms 192.168.64.6

20 76 ms 80 ms 75 ms 192.168.64.77

21 92 ms 78 ms 88 ms www.comcast.net [204.127.195.15]

[prozacnation.livejournal.com]

You realize you sound just like every customer I/this community had yesterday, in the past and also today?

[belovedcrown.livejournal.com]

not exactly. i've worked both sides. i worked various spots from entry level tech support all the way up to director of operations at 2 different isps. I never minded when a customer was upset over an outage- unless they were a residential customer trying to run business services [have no sympathy for you, you should be on business with an sla] or they were overly stupid and/or violent. I am already on business with a guaranteed uptime sla, i'm certainly not stupid [i had to walk the comcast rep through isolating the problem], and i've never been violent or yelled at a tech rep. When you pay good money for a guaranteed service and there are repeated outages due to poor planning and equipment allocation and load balancing by the isp you have every right to be frustrated.

besides, my post here was not simply to whine and vent it was to attempt to further isolate the exact hop that had the problem to provide this feedback to comcast to assist them in troubleshooting.

[belovedcrown.livejournal.com]

1. i am not a dude. i am female. :)
2. the problem was last night [and every other night between 11pm-6am eastern it seems]. that is why you have no trouble loading now.

[shibbychic.livejournal.com]

Member, lurker, now posting.

I know I'm going to get it for this one, but...

I worked for Comcast - or rather, one of the outsourcing companies that was employed by Ccast as a tier one tech agent. We get the same training, so I mind as well have been working for Ccast itself.

What I found was, at the very least, if her call was routed to a call center the my outsourcing company took care of, it is a real possibility that she was talking to a useless, untrained idiot (the kind that make the rest of us look bad).

The tier zero/ tier one techs that worked for that company went way down hill after 90% of the people with skill quit (go figure, abusing your employees will actually make them leave!!) And the ones that stayed? They don't even care anymore, because they're treated so badly, they just powercycle, spend two minuets tops troubleshooting then either roll truck or refer to OEM. Why bother, right? The company treats you like garbage, but they are too short of agents to fire you, there is no room to climb and they are on your back like ebola if your call is longer then 4 - 4.5 minuets.

Now, this company is hardly even training n00b techs (hardly even screeing them, anymore either, I've heard. "Do you have a diploma?" "Duh..." "You're in!")
Since everyone left, they are so hard up for people, there is no choice for them in the matter (had they treated us traind, screen, experienced people well in the first place... oi!). They cut the normally one month training class (3 weeks in class learning, one week Y-jacking [sitting buddy and listening to an experienced agent]) down to a about 2 weeks, and there is little Y-Jacking time. You listen to maybe, four calls, then "g'luck, kid, it's in the manual, but you can't put people on hold to look it up, call times and all. Now move, we're in queue... sucker... we're always in ueue because there's no one call resolution! Mwahahahah!"

Ahem -

Not saying all Ccast and Ccast outsourcing companys have idiot staff, but mind did, and it serviced a lot.

[shibbychic.livejournal.com]

I should clarify, that "Tier Zero" is what we on the phones nicknamed the poor saps who had to take all the Email issues. Because they seemed to get even more stupid people then we got on the phones.
Coincidentally, it was always the people that management hated who got placed here, go figure...

[tjernobyl.livejournal.com]

It seemed that it was a disadvantage to have any actual technical knowledge; if you knew how to fix it, you'd want to go off the script, and that would impair your chances of getting a non-shitty schedule.

[shibbychic.livejournal.com]

Yeah, that exactly. The scripts are another rant all together.