cox is incompetent
-
of course.
after i spend an hour on the phone and demand a site visit, the problem magically goes away
one of my least favorite things debugging telephony circuits back in the day was the call with the tech:
tech: don't see any problem here. can you loop the CSU for me
me: done... see anything
tech: nope.
<lots of typing and tapping on his end. puts me on hold for 10 minutes.>
tech: thanks for waiting. you can drop the loop. i see no errors at all
me: loop dropped. huh. circuit working. what did you do?
tech: nothing. really. honest. didn't touch a thing.uh huh... (wondering who he called and what they did in that 10 minutes)
-
one of my least favorite things debugging telephony circuits back in the day was the call with the tech:
tech: don't see any problem here. can you loop the CSU for me
me: done... see anything
tech: nope.
<lots of typing and tapping on his end. puts me on hold for 10 minutes.>
tech: thanks for waiting. you can drop the loop. i see no errors at all
me: loop dropped. huh. circuit working. what did you do?
tech: nothing. really. honest. didn't touch a thing.uh huh... (wondering who he called and what they did in that 10 minutes)
I used to work telecom, we had a resolution code "PFM".
Half the time it was used was cover for somebody's stupid.
Other times things did clear by Pure Freaking Magic. Sometimes rattling all the connections gets something seated better. Running a high density bit pattern can drive just enough moisture off to get things working again (until next time).
-
I used to work telecom, we had a resolution code "PFM".
Half the time it was used was cover for somebody's stupid.
Other times things did clear by Pure Freaking Magic. Sometimes rattling all the connections gets something seated better. Running a high density bit pattern can drive just enough moisture off to get things working again (until next time).
yeah... i did learn that just running 3 in 24 stress test cleared mystery problems. for a while.
the other bane of my existence (other than backhoes) was techs going into the hole for a new install, finding the pair they'd been assigned was live and just grabbing "an unused pair" and not getting the provisitioning DB updated. sure enough, give it 6 months and they'd double book that pair and drop our circuit.
-
yeah... i did learn that just running 3 in 24 stress test cleared mystery problems. for a while.
the other bane of my existence (other than backhoes) was techs going into the hole for a new install, finding the pair they'd been assigned was live and just grabbing "an unused pair" and not getting the provisitioning DB updated. sure enough, give it 6 months and they'd double book that pair and drop our circuit.
in theory, as a customer, i didn't need to know shit about b8zs, 3 in 24, what a tberd or fireberd test set was, etc. but, the baby bells and sprint both loved going through cycles of firing experienced techs and using contractors, then going back to their own folks but they were wet behind the ears. i do not miss 56k/t1/t3 circuit provisioning and debugging at all.
-
in theory, as a customer, i didn't need to know shit about b8zs, 3 in 24, what a tberd or fireberd test set was, etc. but, the baby bells and sprint both loved going through cycles of firing experienced techs and using contractors, then going back to their own folks but they were wet behind the ears. i do not miss 56k/t1/t3 circuit provisioning and debugging at all.
If you ever had issues with Metro Access or Brooks Fibre in Texas, I'm sorry. My ears were very damp
-
R relay@relay.infosec.exchange shared this topic
-
yeah... i did learn that just running 3 in 24 stress test cleared mystery problems. for a while.
the other bane of my existence (other than backhoes) was techs going into the hole for a new install, finding the pair they'd been assigned was live and just grabbing "an unused pair" and not getting the provisitioning DB updated. sure enough, give it 6 months and they'd double book that pair and drop our circuit.
@paul_ipv6 @johntimaeus @Viss
At my wife's co, back when t1s were a thing, they had all kinds of noise problems. The tech would come out, swap them to the other pair on the run, and leave. Over and over. -
yeah... i did learn that just running 3 in 24 stress test cleared mystery problems. for a while.
the other bane of my existence (other than backhoes) was techs going into the hole for a new install, finding the pair they'd been assigned was live and just grabbing "an unused pair" and not getting the provisitioning DB updated. sure enough, give it 6 months and they'd double book that pair and drop our circuit.
Early in the DSL era, you could remotely put the user modem into hardware loopback (like relay goes click). Then feed straight 48v from the datacenter battery stack for 10-30 seconds. Basically dead short current.
Knocks all kinds if issues right out. Wet mouse condom? It's dry now. And the conductors may have welded together a bit.
You couldn't do it too much in quick repetition, there were a few small fires.
-
Early in the DSL era, you could remotely put the user modem into hardware loopback (like relay goes click). Then feed straight 48v from the datacenter battery stack for 10-30 seconds. Basically dead short current.
Knocks all kinds if issues right out. Wet mouse condom? It's dry now. And the conductors may have welded together a bit.
You couldn't do it too much in quick repetition, there were a few small fires.
@johntimaeus @paul_ipv6 im 100% sure this is some fuckery on their part. i downgraded from gig to 500mbit metered because it was less than half the price, and almost immediately afterwards i started getting multiple 2-5 minute outages per day. its gotta be some bullshit logical something or other on their end
also, i wouldnt put it past them to introduce deliberate problems just to punish people who decided to give them less money, because its nearly impossible to prove from this end
-
@paul_ipv6 @johntimaeus @Viss
At my wife's co, back when t1s were a thing, they had all kinds of noise problems. The tech would come out, swap them to the other pair on the run, and leave. Over and over.Very few people learned how to hook an oscilloscope up to a dry pair and look for noise before deciding which one to run on.
-
Very few people learned how to hook an oscilloscope up to a dry pair and look for noise before deciding which one to run on.
@johntimaeus @paul_ipv6 @Viss
Well I mean they were both bad. They just didn't want to run a new line. -
If you ever had issues with Metro Access or Brooks Fibre in Texas, I'm sorry. My ears were very damp
sprint and us west/worst were the real banes of my existence.
-
@johntimaeus @paul_ipv6 im 100% sure this is some fuckery on their part. i downgraded from gig to 500mbit metered because it was less than half the price, and almost immediately afterwards i started getting multiple 2-5 minute outages per day. its gotta be some bullshit logical something or other on their end
also, i wouldnt put it past them to introduce deliberate problems just to punish people who decided to give them less money, because its nearly impossible to prove from this end
Possibly a traffic shaper doing a log or database sync?
-
sprint and us west/worst were the real banes of my existence.
@paul_ipv6 USW was horrible. I had a backup OC48 through them that couldn't achieve one 9 reliability.
-
Possibly a traffic shaper doing a log or database sync?
@johntimaeus @Viss @paul_ipv6 nope. It's Cock's "fairness" policy. They intentionally oversell to the breaking point everywhere. Where they can legally offer 'metered' bullshit, they intentionally shove all of those customers into extremely low commit streams so that the unlimited customers can easily squeeze them out. Where they can't do metered bullshit, they just oversell well past the breaking point and tell you to get Internet from someone else then.
-
@johntimaeus @paul_ipv6 @Viss
Well I mean they were both bad. They just didn't want to run a new line.I used to have a three page cheatsheet on getting around problems like that. None were good, but they worked.
I got it from an engineer at Fujitsu. Crazy stuff like using 2 pair with an extra loop on one to carry one pair's worth of signal. It schmears the transition, but still gets the job done at 1.5 mhz.
