Today Was Different – AT&T Hangs

For about a week at just a little after 1 PM, our internet connection would hang. Pinging inside the network showed all my gear up and fine. The boundary router (AT&T router) up and running. Just the AT&T router to AT&T site, no-go.

At that point I’d do a power-fail, reset, on the router, it would reboot, and all would continue as normal again.

I’ve not done more debugging that that. I can see the status page on the router in a browser, so maybe tomorrow I’ll query it some more.

But today was different. It hung at about 12:22, just after noon.

I’d figured it was just some work being done at the AT&T site and it would all end “soon”. Now I’m not so sure. It had seemed like someone returning from lunch, powering up some network gear and causing their router to hiccup. But at just a bit after noon? Now that thesis is toast.

Sigh.

Not a horrible thing. But really. Were I doing a day long download of something, or any kind of bits transfer of size really, I’d be royally pissed.

I sure hope this isn’t some kind of “new policy” and is just a local screw-up.

Whatever it is, AT&T is slowly convincing me that other ISPs will be more reliable. (It’s hard to be worse than “at most 24 hour uptime”… )

Subscribe to feed

Advertisements

About E.M.Smith

A technical managerial sort interested in things from Stonehenge to computer science. My present "hot buttons' are the mythology of Climate Change and ancient metrology; but things change...
This entry was posted in Tech Bits and tagged , , , . Bookmark the permalink.

11 Responses to Today Was Different – AT&T Hangs

  1. E.M.Smith says:

    Well this isn’t good:

    Timed Statistics
     	15 Min 	Cur Day 	Showtime 	Last Showtime 	Total
    Errored Seconds (ES) Line 1 	144 	197 	187 	197 	197
    Errored Seconds (ES) Line 2 	0 	10 	0 	0 	10
    Severely Errored Seconds (SESL) Line 1 	16 	46 	36 	46 	46
    Severely Errored Seconds (SESL) Line 2 	0 	10 	0 	0 	10
    Unavailable Seconds (UASL) Line 1 	0 	128 	0 	0 	128
    Unavailable Seconds (UASL) Line 2 	0 	118 	0 	0 	118
    FEC Errors Line 1 	1101760 	1811521 	1825478 	1825500 	1811521
    FEC Errors Line 2 	0 	0 	0 	0 	0
    CRC Errors Line 1 	4894 	13397 	9672 	13197 	13397
    CRC Errors Line 2 	0 	6487 	0 	0 	6487
    DSL Initialization Timeouts Line 1 	0 	0 	0 	0 	0
    DSL Initialization Timeouts Line 2 	0 	0 	0 	0 	0
    

    Now I’m wondering if it’s a problem with thermal dependent errors on the line / link…

  2. ossqss says:

    @EM, have you dug into the ATT router settings? There may be some options in there to help. I had an issue a while back with DDOS (Spectrum) and flipped that protection on and things improved. I still reboot periodically to switch IP address as it is not static. Not saying that is your case, just thought there may be some error correction tools available in there. Most times I have had intermittent service interuption has been a power supply or amp issue at one of the local pods upstream. Tech network support was able to see it when I called them. Good luck!

  3. John F. Hultquist says:

    A few years ago we had a lot more trouble than we do now.
    We are on the phone company’s lines. They replaced the router at one point, claiming the one we had always developed issues when copper lines were involved. They have also upgraded things out our way. I think we are now on a loop rather than a dead-end of DSL.
    We do have drop-outs now and then — I think that is in-house. A few clicks or a 1/2 minute wait, or both, usually fixes it.

    Yesterday morning (first in a couple of months) the internet light glowed red.
    I learned to count to 15 many years ago so have gotten good at pulling the cord.
    This time it worked as described.

  4. ossqss says:

    @EM, I looked and it appears there are some power issues in Cali. You may be a victim of upstream equipment power issues. Some kind of Flex alert thing? I will not comment further on solar and wind sources…..

  5. E.M.Smith says:

    @Ossqss:

    Oh, good point… I’m on a small UPS and the Telco nodes out here are on some kind of bigger UPS PDU thing (big box end of the block). Don’t know about further upstream or what kind of “shedding” agreements they might have. I could see things “glitching” as they do a “peak sun” controlling activity… About 1 PM middle of summer is just about Peak Sun, so all sorts of non-sun gear will be shutting down at that point and who knows what happens to frequency stability and noise pulses / volts sags.

    The Duck Curve gets a bit savage about 7 PM when winds die down and solar is history and almost EVERYONE is home from work firing up the AC and cooking and TV and…

    IIRC there’s been a statement that So Cal especially was on some kind of power alert and to please not use electricity right when you need it the most…

    Would also fit the onset of the behaviour a few weeks ago right at peak sun and heat onset.

    I could see saggy volts and surges / drops / pulses & clicks causing high perf coms gear to be annoyed and glitch. It’s rock solid at night…

    I wonder if crappy power will first show up as crappy internet communications…

    FWIW, I’d been just hoping it was temp work and would just go away. Only know have I decided it’s likely up to me to fix or escalate it. So no, I’ve not gone into the guts of it and settings yet. ( I usually get paid to debug network issues ;-)

    @John F.H.:

    Well, I’ve put the house AT&T router on a radio controlled switch so I can just use a clicker to do a power fail / restart on it. We use that to shut it off each night (so sleep is better – since we found the WiFi is part of the sleep disruption issues. Killing 5 GHz helped the most, but box off at night is also detectable improvement.)

    All the LEDs are all green all the time; despite the log file (posted above) showing errors.

    It’s possible the wires are just crappy. AT&T doesn’t seem to bother making sure the wire is good. Just replaces it if things get so bad they fail entirely. Had that happen when they first put this line in about 2 years ago. Prior to that, windy days caused lots of line noise. Now we don’t have a phone with them so no idea on line quality upstream.

    OTOH, it’s a once a day thing pretty much every day at about noon-1:30 so fairly easy to plan around and a reboot only takes a minute and fixes it… so my motivation has been marginal.

  6. Larry Ledwick says:

    Years ago I had a similar issue with glitches, after a bit of observation I realized those glitches matched up to when near by industrial plants would be shutting down for lunch breaks and end of shift or first thing in the morning starting all their motors, causing a sudden load on the system and screwing up load factors and probably generating inductive spikes. That is why I bought my first UPS for my home computer. Once I got that the issues went away for me.

    I wonder if you have a similar issue due to switching loads or shedding loads at certain times of the day.

    Now days unless the power drops completely I never even notice. I have surge protectors at the wall sockets, then power for the computer goes through a hospital quality isolation transformer (inductive coupling only, no direct current path to the load side) then that feeds a good quality Tripp Lite UPS. Even when lightning is cracking nearby I rarely hear it beep due to a momentary power sag or cut out.

  7. E.M.Smith says:

    @Larry:

    I know my power has dropouts and surges as the light bulb flickers sometimes…

    But the monitor doesn’t nor does the TV.

    All the computer gear including the AT&T router are on a UPS and stay up. What has an issue is the link on the Telco side – so their gear outside the house somewhere.

    Or, possibly, line noise and quality issues cause errors to accumulate to a freeze point for that DLS line and a reboot clears it…

  8. Larry Ledwick says:

    Understand, I was having similar problems with comcast for quite some time, every now and then the data rate would just go to a crawl, after much head scratching I found that full power reboot of the modem and router made the problem go away.

    They must have fixed it recently because it has been 6 weeks or more since I last had to do that but when it happened it would make my vpn to work just hang sometimes taking several minutes to echo a command to the screen. Nothing wrong with the vpn itself as others I was working with would be having no latency issues at all, but my connection was completely unusable.

    It could take me 15 – 20 minutes to do a task that should take about 30 seconds.

    I feel your pain – drove me nuts for a long time until I figured out how to fix it.

  9. E.M.Smith says:

    Well the good thing is that it only happens once or twice a day and is trivially fixed with an AT&T box reboot. Only the one box from the Telco. Has DSL & boundary router in one skin.

    also only happens in a narrow time window. Quite livable really.

  10. p.g.sharrow says:

    Yeah an ATT glich. We were having a similar problem In Chico. The feed would slow to a stop morning and nite. A reboot generally would bring it back up. After many complaints ATT would replace their box and things would work fine for awhile. I am told that they buy the cheapest Chinese made boxes. and the failure rate is very high. Also they tend to only fix things after a total failure or constant system overload. They do constantly slip in extra charges, upgrades and fees so you best check every bill before you pay them.
    ‘I would suspect they are on the edge of cash flow disaster in servicing all their financing of acquisitions. Acquisitions made to justify refinancing and increase cash flow. Normal corporate congame to enrich top management with golden parachutes…pg

  11. beng135 says:

    My DSL (land-line) connection goes out in rainy weather. Turn the modem off-on, and it recovers for a minute or two, then back out again. Not sure why — perhaps there’s a satellite or cell tower connection in the chain somewhere…..

Comments are closed.