[SOLVED] I know it could be a bunch of reasons, but what do you think could be causing me random 10-60s disconnects about 2-3 times daily?

omonoiatis9

Distinguished
Sep 23, 2013
75
0
18,630
The router (provided by my ISP) is an ARRIS TG2482 and I have 3 PCs and 1 smart tv directly connected to it with ethernet cables. For the last month and a half, the entire network (all pcs/tv) lose internet connection completely, for anywhere between 10 seconds to 60 seconds before it gets restored, almost daily. Sometimes it's 1 time, sometimes it's 2-3 times, sometimes if I'm lucky a day goes by with no disconnects at all.

What are some common reasons off the top of your head that could cause something like this? Some things I could remove through method of elimination:

a) It's not the cables, cause we're talking about 4 different devices with 4 different cables.
b) It's not the PCs, cause again, we're talking about 4 different devices and all of them lose connection at the same time, and are directly connected to the router.
c) The ISP says (though I don't believe them) that it's probably not the router, cause they checked from their side and everything seems normal with it - did a factory reset too. I welcome you to challenge this claim. I read that the operating temperature of this specific router is up to 40C, but touching it burns my hand (obviously a lot hotter than 40C and closer to 60C if i had to guess). Could an overheating router cause random network dropouts so few times per day?
d) It's not some wireless device overloading the network or something, there are no WIFI devices connected at all.

Extra info that might be relevant:

-I've been with this ISP for like a decade and they've been very stable. This issue only started about a month or two ago.
-According to my ISP's website, the tech they use to provide internet is DOCSIS 3.0 . Any network engineers can point to a potential flaw of the setup in the electric pole that could cause these issues? I remember about a month and a half ago (around the time these issues started happening - also around the time they gave me a new router), there were network engineers from my ISP outside my building working on the pole for hours for 2-3 days straight.
-I mentioned all of the information that i'm giving you in this post to my ISP already, and they said they'll send engineers next week to look into it, but I wanna be a bit more informed and see what you guys think could be causing this.

I'm convinced it's either the router that's broken or they messed something up during their maintenance work on the pole 1-2 months ago, cause these 2 events are the only events I can think of that coincide with when this issue started.
 
Last edited:
Solution
Your best first place to look is in the router logs. You might see error messages around the time you are getting the outage.

Otherwise it is the standard test method. Run tracert to 8.8.8.8. It will likely not show any issues, your goal is to get IP addresses of devices in the path.

Now open at least 3 cmd windows. Run a constant ping to hop 1 (your router) hop 2 (the first ISP router) and 8.8.8.8.

What is the most common is you will see no loss to hop 1 but see loss to hop 2 and 8.8.8.8 when you have issues. This means there is some problem with the connection between your house and the ISP. You might already have seen this in the messages if you check the log.

If you get issue to hop 1 it is some issue...
Your best first place to look is in the router logs. You might see error messages around the time you are getting the outage.

Otherwise it is the standard test method. Run tracert to 8.8.8.8. It will likely not show any issues, your goal is to get IP addresses of devices in the path.

Now open at least 3 cmd windows. Run a constant ping to hop 1 (your router) hop 2 (the first ISP router) and 8.8.8.8.

What is the most common is you will see no loss to hop 1 but see loss to hop 2 and 8.8.8.8 when you have issues. This means there is some problem with the connection between your house and the ISP. You might already have seen this in the messages if you check the log.

If you get issue to hop 1 it is some issue with the router, If it is only 8.8.8.8 then it is some error farther into the ISP network. If they all show no outage when you think the internet is down it is going to be more complex.
 
Solution

omonoiatis9

Distinguished
Sep 23, 2013
75
0
18,630
Thanks for your input @bill001g

I checked my router's system logs and they do show some interesting errors (that I have no idea how to interpret). Router log . Here is also upstream/downstream power levels What do they mean? A google search tells me that event 3 means critical error. So the router is the culprit since it does have quite a few event 3 errors? Also, isn't my ISP also able to look at my router's logs from their side? Are they that incompetent that I called them 2 times to troubleshoot this thing and they never bothered to check logs or at least ask me to check them?

Also, I just ran some tracert. 2nd hop is always timing out. Is this normal? I see 4 IPs that belong to my ISP, 2 that belong to google. So what you suggest is to run a 24/7 ping to a) My router (192.168.0.1) b) First ISP Ip that appeared in my tracert and c) 8.8.8.8, then compare the logs and see where they overlap?

What if I get issue with both hop 1 and hop 2? Does that still mean it's the router?
 
Last edited:
The ISP for whatever reason is blocking ping to their router ip. Make troubleshooting harder. You would find the first hop that works but that leaves the ISP to claim stupid stuff like there routers show no errors because you can't prove otherwise.

You almost never see problems to hop1. You most times can unplug the internet connection and it will still ping hop1. Either the router is somehow failing or you have a ip conflict. If for example another device in your house is claiming the 192.168.0.1. If a second router were hooked up that would cause it.

It depends how often your problem occurs. If possible leave a ping run to the router IP with only 1 PC plugged in. If it has issues doing that you really start to suspect the router. If that works add devices 1 at time until maybe you see the problem.

Can you get lucky and see the lights on the router when it does this. Maybe the router is rebooting....that you should see in the log. With the problem being in hop1 it likely doesn't matter but I would check the power levels on the modem part of your router. You should also be able to see number of errors. You will always see some but these numbers should not increase greatly. Checking this stuff is to see if maybe there are line errors but normally you would see hop 1 good and errors on hop 2....or whatever the next hop is. When you have errors in hop 1 it contaminates any ability to test past the router.
 

omonoiatis9

Distinguished
Sep 23, 2013
75
0
18,630
The ISP for whatever reason is blocking ping to their router ip. Make troubleshooting harder. You would find the first hop that works but that leaves the ISP to claim stupid stuff like there routers show no errors because you can't prove otherwise.

You almost never see problems to hop1. You most times can unplug the internet connection and it will still ping hop1. Either the router is somehow failing or you have a ip conflict. If for example another device in your house is claiming the 192.168.0.1. If a second router were hooked up that would cause it.

It depends how often your problem occurs. If possible leave a ping run to the router IP with only 1 PC plugged in. If it has issues doing that you really start to suspect the router. If that works add devices 1 at time until maybe you see the problem.

Can you get lucky and see the lights on the router when it does this. Maybe the router is rebooting....that you should see in the log. With the problem being in hop1 it likely doesn't matter but I would check the power levels on the modem part of your router. You should also be able to see number of errors. You will always see some but these numbers should not increase greatly. Checking this stuff is to see if maybe there are line errors but normally you would see hop 1 good and errors on hop 2....or whatever the next hop is. When you have errors in hop 1 it contaminates any ability to test past the router.

I also included upstream/downstream power levels if that can tell you anything. The screenshot is from 6 hours of uptime. I'll try to check the router's lights when it's happening to see if it's restarting but I doubt it, considering that the connection needs more than a minute to come back after a router reboot, and the disconnects usually last less than a minute.

I can't really leave only 1 pc on the router unfortunately. There's always at least 2 in use. But still, if it was another device trying to claim an already used IP, wouldn't this mean that only 1 other device would get disconnected (the device whose IP is being claimed?). All devices get disconnected every time this happens. From the few google searches I did people seem to point out that T3/T4 timeouts and connection interruptions are pretty much always going to be faulty signal somewhere which means a physical connection problem . Either in the house or outside the house. So maybe the last tech that came to install this new router 1-2 months ago messed something up with the splitter or used a bad splitter. Or maybe the connection from the pole to the building is bad somewhere. Hopefully they send a competent tech that can figure it out.
 
Last edited:
The levels look very good. The number of errors depends on how long it has been since that was cleared. It is a big problem if it is only a hour or two but if it is many days or even weeks they are fine.

If you get t3/t4 timeouts or even worse sync errors in your modem log there is a big problem. Most times you get a few error messages here and there over a couple days.

Your levels though are very good. Many times the problem will be you see very high upload power levels. This means the ISP is having trouble haring your modem and asked it to transmit at more power. Your numbers are almost perfect in the low 40's. Problem ones are near or over 50.

But I suspect this is going down the wrong path. There is something wrong in the lan that needs to be solved first. I leave ping commands run to my router and other IP addresses constantly so if I get stalls I can check. I will lose 1 ping out of say 10,000 to the router.

What I was thinking was maybe something claimed the IP of the router. That would break all the machines and it would be kinda random because of how ARP works.

Maybe another test would be to leave a constant ping run both to the router IP and to another machine on your network. The lan ports on a router are a simple switch. Pretty much even if the router cpu get hung up it will still pass data between lan ports because this is more of a hardware function than software. If you get good ping between machine but not to the router IP either the router itself has a problem or again the IP is somehow wrong.
 

omonoiatis9

Distinguished
Sep 23, 2013
75
0
18,630
The levels look very good. The number of errors depends on how long it has been since that was cleared. It is a big problem if it is only a hour or two but if it is many days or even weeks they are fine.

If you get t3/t4 timeouts or even worse sync errors in your modem log there is a big problem. Most times you get a few error messages here and there over a couple days.

Your levels though are very good. Many times the problem will be you see very high upload power levels. This means the ISP is having trouble haring your modem and asked it to transmit at more power. Your numbers are almost perfect in the low 40's. Problem ones are near or over 50.

But I suspect this is going down the wrong path. There is something wrong in the lan that needs to be solved first. I leave ping commands run to my router and other IP addresses constantly so if I get stalls I can check. I will lose 1 ping out of say 10,000 to the router.

What I was thinking was maybe something claimed the IP of the router. That would break all the machines and it would be kinda random because of how ARP works.

Maybe another test would be to leave a constant ping run both to the router IP and to another machine on your network. The lan ports on a router are a simple switch. Pretty much even if the router cpu get hung up it will still pass data between lan ports because this is more of a hardware function than software. If you get good ping between machine but not to the router IP either the router itself has a problem or again the IP is somehow wrong.

All the errors and uncorrecteds shown in the screenshots above are from just 6 hours of uptime. That's not normal is it? And just this morning new errors appeared in the log:

7/9/2021 8:38​
82000400​
3​
Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 8:42​
82000200​
3​
No Ranging Response received - T3 time-out;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:06​
84000700​
5​
RCS Partial Service;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:06​
84000500​
3​
SYNC Timing Synchronization failure - Loss of Sync;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:06​
84000700​
5​
RCS Partial Service;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:06​
84000500​
3​
SYNC Timing Synchronization failure - Loss of Sync;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:06​
84020200​
5​
Lost MDD Timeout;CM-MAC=censored;CMTS-MAC=censoredCM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:06​
82000200​
3​
No Ranging Response received - T3 time-out;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:35​
84000700​
5​
RCS Partial Service;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​
7/9/2021 9:39​
82000200​
3​
No Ranging Response received - T3 time-out;CM-MAC=censored;CMTS-MAC=censored;CM-QOS=1.1;CM-VER=3.0;​

I also went to the pings that i was logging overnight at the time these errors occurred. Every time there was a "SYNC Timing Synchronization failure" there were constant "request timed out" messages for about 1 minute on all 3 of my pings (ping to 192.168.0.1, ping to some IP at my ISP, ping to google DNS). On the "No ranging response received" the ping returned "General failure". But here's the really weird one: During one of those "SYNC Timing Sync Failure" events (which lasted about 4 minutes - one of the longest ones), the ping to my router DOES give responses, and the ping to ISP and google DNS give timeouts. Does this say anything?

Additional info: I have 4 machines total connected to the router. 1 smart TV and 3 desktop PCs. But the smart TV and 1 of the PCs are almost always turned off, so in total just 2 PCs. I tried pinging the other active PC but I constantly get "request timed out" no matter what. Is this normal? As a test I told my brother to change his IP from dynamic to static while I keep mine on dynamic to see if that helps but I doubt it.
 
Last edited:
Sync errors are of the worst kind you can get. Many of these cause your modem to go into its retrain mode which can take many minutes. Sync errors are almost always a issue with the ISP equipment on the remote end. Call the ISP they should be able to see the messages in your modem....many times this stuff is logged on the ISP equipment also.

This gets hard when you might also have issues in your lan. Do not tell the ISP anything about ping loss to the router. Just show them the logs and complain about network slow and not working. When it is working well you should see only a few messages a month in this log.