New Member
•
30 Messages
DSL Service is Flapping
I've had the AT&T DSL service (25Mb/s tier) for several years. It's been rock solid. My significant other and I both work from home and have never been in danger of hitting the hard limits. We previously had cable that promised some pretty substantial throughput (upwards of 10Gb/s) which sounds awesome, except for the fact that the service was shared by everyone in the community (partyline, essentially), and had a really bad track record of going out during the day, which meant 0 Mb/s, until they got around to fixing it.
We switched to AT&T ADSL Service at 25 Mb/s, which is dedicated and shared with no one. The service has been reliable... until very recently. One evening about a month ago, suddenly all of the services in the house stopped working. No internet browsing, no streaming (television content), no making cell phone calls over WiFi (WiFi service (Edited per community guidelines), so using cell over WiFi)... to name a few. After performing a bunch of diagnostics on the ARRIS device, my theory was that either the device was malfunctioning, or there was a hardware problem somewhere upstream. I reached out to AT&T Cust Service. The gentleman on the phone spent most of the call handling me, which I'm fine with if he is ultimately able to escalate to an engineer. He spent an inordinate amount of time interrogating the device in my home (the ARRIS BGW210-700), I spent about an hour on the phone while he alleged to be running tests, he was making the case that the problem was between the ARRIS and my home network. I explained repeatedly that I had already run the diagnostics on the ARRIS and that they clearly showed that the problem was upstream. Here are some images of those diags:
Figure 1: Running the internet speed test using the ARRIS-based utility complained because it couldn't contact the Test Server, and as a bonus, notice the massive latency, 8758ms:
Figure 2: After running the suite of ARRIS on-board diagnostics, the IP test failed, indicating a possible routing hardware issue:
Figure 3: Clicking into the details of the IP Test failure yields more insight into the problem, related to access to the Default Gateway (a problem upstream from my equipment):
Figure 4: Finally, a "PING" test to google.com which was coming back with 100% packet loss, in this case showed 50% packet loss plus the latency issue.
The customer service rep eventually agreed to change out my modem, which I installed two weeks ago. The screen shots in this post were captured last evening, evidenced by the date/time stamps. The new hardware didn't fix the problem.
Would it be possible to speak with an AT&T engineer to figure out why this is happening. This morning, as I write, all tests indicate everything is functioning properly, but I suspect that whatever caused the malaise yesterday afternoon (and a few weeks go) and lasted into the evening hours (flapping service) will likely happen again. If there was a service outage due to routine maintenance, I was hoping to learn that from customer service, but customer service didn't share that when I called yesterday.
Are there any engineers on this forum? I'm hoping this detail will help bypass the customer service support tier.
gz1
New Member
•
34 Messages
7 months ago
With all due respect, it has not been this way for years - for everyone posting in these two threads it's been this way since early February. What we're seeing is more than just a lousy connection, we're seeing our internet connections completely shutting down to the point where even pings stop. The router should have mechanisms in place when upload speeds get near their saturation point to let other traffic work, and in fact it did until recently.
Can one's Internet be a little pokey and cause latency when you are doing something with lots of upstream? Sure
it can run a little slow or cause some buffering in streaming video and the like. But not like we've been experiencing since early February.
We've all had this same service for years, doing the exact same workloads and none of this took place until a few weeks ago. Doing an offsite backup should not take my internet down for hours while it runs. It never did until early February and this backup has been running every day for years.
Can I find solutions myself to limit upload bandwidth if this persists? More than likely yes, but it will be a total hassle and should not be necessary. Hopefully the same firmware process that caused the issue on the AT&T router can fix it if we can get the word out.
(edited)
0
0
JefferMC
ACE - Expert
•
33.2K Messages
7 months ago
Maybe it should. The feature is called QoS and a lot of third party routers have it. But historically the Gateway has only prioritized VOIP over U-verse TV, and both of them over regular Internet traffic.
I don't know what to say. I've seen this happen dozens and dozens of times. Perhaps somebody at Microsoft tweaked a Windows setting for IP window management and made things worse. Maybe AT&T tweaked something to make upstream faster for fiber. Perhaps something has been added to your traffic load that you aren't aware of are or are just not thinking about. Maybe you've got malware.
First happened to me when my daughter came home from college and her phone got on the home Wi-Fi and decided to backup everything to Google. For 90 minutes I hunted for the problem before finding the troublemaker. I then used my third party router to limit the bandwidth for that device, as that was the best that particular router would do at that time.
0
0
gz1
New Member
•
34 Messages
7 months ago
Agreed completely, these problems can be really tricky to track down and for any one person it could be a large number of things. That's why we're crowdsourcing our problems on these threads, to find similarities and eliminate things via shared information. Given how many of us have almost the exact same symptoms starting from almost the exact same date, I'm more and more confident it's not going to be anything unique on our internal networks that is causing this.
An AT&T change is what I'm currently convinced it is, hopefully with more sharing of info we can get more specifics on exactly what. Heck, even if I'm off on all this, we're one step closer...
(edited)
0
0
JefferMC
ACE - Expert
•
33.2K Messages
7 months ago
Does everybody involved in both of these threads have low capacity DSL vs having Fiber?
0
0
gz1
New Member
•
34 Messages
7 months ago
It's looking like that, in the other thread people are starting to list their connections and I think everyone so far is DSL.
I pretty sure one person on DSL Reports mentioned a similar issue in the thread discussing the firmware upgrade being on fiber, but there wasn't a lot of details.
0
0
mag2023
New Member
•
30 Messages
7 months ago
@JefferMC ... It seems that way. My DSL service is the 25/5 residential service. No fiber available in my community... yet.
0
0
JefferMC
ACE - Expert
•
33.2K Messages
7 months ago
I just wanted to be sure we did have the same connection information. One thing about the symmetrical nature of the fiber connection is that it normally alleviates this problem; in fact QoS features of third party routers usually only slow down traffic on fiber and are thus counter productive.
(edited)
0
0
mag2023
New Member
•
30 Messages
7 months ago
Posted in the other thread last night, but wanted to re-post here since I started this thread....
Hey All. I've got an update. The customer service representative that I was working with on Friday evening called me back last night to check in on things. I told him that I met a number of AT&T DSL customers on the AT&T forums on Saturday that have the same problem and are running the same equipment. I told him that one of our theories is that a firmware update (the dreaded 4.21.5 update) in the Jan/Feb timeframe causes device instability. I also explained that I was able to precipitate the problem by hammering the upload.
He asked me if I'd like to try swapping the ARRIS for a different modem model (the PACE 5268AC). It's arriving on Wednesday, so I'll let you all know how things go.
stay tuned....
0
mag2023
New Member
•
30 Messages
7 months ago
Hi Folks,
The Pace 5268AC arrived after dinner and I just swapped it in for the ARRIS BGW210-700. After the installation process completed (about 10 minutes), it came online. Speed tests were right in line with what I saw with the ARRIS... 30dn/5up. My pings are solid.
I was able to knock over the ARRIS by uploading a large file, which I did using a 2GB fedora iso. I kicked off the upload and while the throughput and ping times took a hit, there were no ping timeouts -- That's good news! Fingers crossed, the Pace will behave going forward.
0
0
gr8sho
ACE - Professor
•
4.5K Messages
7 months ago
Users are reporting the 5268 will no longer receive firmware updates, which presumably is a security worry, but in the case of that gateway is probably a good thing!
0
0