nine fives of service
Nov. 3rd, 2021 10:09 pm[twitter, again]
Remember this?
It's Wednesday 10pm. We're _still_ still broken.
The last person to call us was a Level 1 support tech who insisted we power-cycle a modem which has a built-in multi-hour backup battery.
I got a screwdriver and made it happen. It didn't help.
(@comcastbusiness)
We're getting told contradictory things about the status of our ticket, we're getting told (by "power cycle the modem with the built-in battery") that SecurityEdge is absolutely disabled but maybe it is and maybe it isn't, I've gone in DM and said "I want these things exactly."
Right now I'm at the point where I have no idea whether Tier 2 who got it fixed the first time by turning off SecurityEdge and said this was a known issue was actually telling us the truth or not.
I'm genuinely in a state where _I just don't fucking know_.
But what I do know is that direct requests from our name server to root servers are getting DNS format error/non-improving referral errors. But if we do it _through_ the same server from a _client_, we _don't_.
Still.
And the only cases I can find with searches where this shows up are:
1. The remote DNS server is misconfigured (_unlikely_, given it's the _root servers_), and,
2. The upstream ISP fucking up the routing and/or packets, causing this.
Guess what I'm pretty sure is happening?
I know some IT people who joke that @comcastbusiness strives for "nine fives" of service. (Vs. five nines, right? 99.999% uptime.)
We are now actually below nine fives of service, hobbling along stealing DNS service from G**gle and Comcast DHCP's DNS servers, and I am screaming.
Remember this?
It's Wednesday 10pm. We're _still_ still broken.
The last person to call us was a Level 1 support tech who insisted we power-cycle a modem which has a built-in multi-hour backup battery.
I got a screwdriver and made it happen. It didn't help.
(@comcastbusiness)
We're getting told contradictory things about the status of our ticket, we're getting told (by "power cycle the modem with the built-in battery") that SecurityEdge is absolutely disabled but maybe it is and maybe it isn't, I've gone in DM and said "I want these things exactly."
Right now I'm at the point where I have no idea whether Tier 2 who got it fixed the first time by turning off SecurityEdge and said this was a known issue was actually telling us the truth or not.
I'm genuinely in a state where _I just don't fucking know_.
But what I do know is that direct requests from our name server to root servers are getting DNS format error/non-improving referral errors. But if we do it _through_ the same server from a _client_, we _don't_.
Still.
And the only cases I can find with searches where this shows up are:
1. The remote DNS server is misconfigured (_unlikely_, given it's the _root servers_), and,
2. The upstream ISP fucking up the routing and/or packets, causing this.
Guess what I'm pretty sure is happening?
I know some IT people who joke that @comcastbusiness strives for "nine fives" of service. (Vs. five nines, right? 99.999% uptime.)
We are now actually below nine fives of service, hobbling along stealing DNS service from G**gle and Comcast DHCP's DNS servers, and I am screaming.
no subject
Date: 2021-11-04 06:24 am (UTC)Looks like whoever tried that at ComCast did not consider that the design challenge is supposed to be mental masturbation for systems nerds, not something you actually IMPLEMENT!
Also, that sucks, have whatever comfort you can from my commiserations.