How to troubleshoot intermittent drops in my internet connection?

February 22, 2013 1:24 PM

How to troubleshoot intermittent drops in my internet connection?

Hi! I have a periodic problem with my internet connection that I am trying to troubleshoot.
In general, my internet works fine. But every so often, it disappears for a few minutes. It does so in ways that seem weird to me, and I'm trying to figure out. The problem is hard to troubleshoot in part because it is so unpredictable and intermittent.

I would be very grateful for:

1) Any tips as to what the problem might be
2) Any tips on tools I could use to help diagnose or troubleshoot the problem.


-----

Here is my best possible description of things:

** My level of expertise/understanding:
I am a computer-savvy person who does not have a ton of network expertise. My level of expertise = I know superficially what DNS is, what ping is, that sort of thing. Was able to set up my own home network on Tomato w. some effort. But there is lots I don't know.

** Network setup:
- My internet connection is w. Teksavvy in Toronto. It is a 25Mbps DSL connection
- The net connection goes through a WRT-54gl router running Tomato. This connects to three computers over ethernet (all running Windows 7), as well as the occasional laptop, printer, and portable device over WiFi
- I use unblock-us.com as my DNS on all of the computers running windows 7. The router is not configured to use unblock-us.com.
- Most of the time, as near as I can tell, the network runs really well. Speedtest tells me I am getting the speeds I should be getting, ping times are low, etc.

** The problem
- Periodically, it seems that none of the machines on the network can load any web pages. When this problem happens, it seems to affect all machines simultaneously. No computer on the network can load web pages or check email. The problem seems to affect all computers and all installed browsers (firefox, explorer, chrome)
- *However* - If I open a command prompt in windows, and try to ping a remote computer, it generally works: I can ping facebook.com, yahoo.com, etc.. So I am not totally off the internet. But it seems I can ping some sites and not others. (In one recent outage: I could ping facebook and yahoo, but not google...)
- If I log into the Tomato admin panel for my router ,there's a ping tool there. Using that tool- when the problem is occurring, it seems that I am unable to ping any web site if I specify it by domain name. But it seems I am able to ping at least web sites if I specify them by IP number. (eg: I can ping 8.8.8.8) (And, when I am not having this problem, the router can ping web sites by domain with no problem).
- Sometimes when this happens, I try running Windows' network troubleshooter. It seems to give different results at different times. Most recently it told me "Windows tested your internet connection and verified that you are able to access some websites." which suggests that it cannot see a problem. But not of the computers on my network can load web sites using any browser.
- In general, while I'm not 100% sure, it seems to me that the problem goes away on its own after 5-15 minutes. I usually try restarting the modem, the router, the computers. But I have not discovered any series of restarts that automatically fixes the problem, and I'm pretty sure taht at least a couple of times I've just let everything sit for a few minutes and then internet service is restored.

* Troubleshooting challenges
Apart from my own limited knowledge, one thing that makes troubleshooting hard is how intermittent and brief the problem is. So, for instance, someone from my ISP suggesting plugging my computer directly into the modem, to isolate the router from the problem. But to do this, I'd potentially need to disconnect two of my computers from the internet for a couple of weeks, which isn't practical. Similarly- when the problem occurs, I 'd love to be able to do whatever testing/troubleshooting I can as quickly as possible...

posted by ManInSuit to computers & internet (16 answers total)

Have you simply rebooted the router?

What you describe is not uncommon for consumer-level routers, especially the WRT-54gl. People I know who have that router are rebooting (not resetting) them all the time. It's just the nature of it, as far as I understand.
posted by TinWhistle at 1:33 PM on February 22 [1 favorite]

Not sure if this will help, but have you looked at the issue using inSSIDer? I was having some trouble with my wireless and the kind folks at AskMe turned me on to this tool.
posted by blurker at 1:39 PM on February 22 [1 favorite]

Yes, this is one of the reasons I finally gave up on Linksys routers. I had a WRT54GL that did this all the time. I installed Tomato because everyone claimed it would fix the problem, but it didn't. I'm pretty sure the problem is that the hardware actually overheats. Rebooting it typically would fix it for me, but having to do that got really old, so I replaced it with something more robust and haven't had a single problem since.
posted by primethyme at 1:40 PM on February 22

Thanks TinWhistle! A bunch more questions, then:

1) What's the difference between rebooting and restarting the router?

2) Would rebooting the router periodically help prevent this problem? Or is it just what I need to do when the problem occurs?

3) Is it that common? I get outages like once or twice a week. That seems like a lot! Is that just to be expected? Or a sign that my router is old and needs to be replaced? If I replaced this router with a new one of the same model, would the problem likely go away?

4) Is there anything I can to do tell whether the problem is indeed with the router?
posted by ManInSuit at 1:41 PM on February 22

blurker - it looks like inSSIDer is for troubleshooting wireless networks. Is that right? My problem is with a wired network.

primethyme - Oh! Interesting to hear more reports of router woes. Maybe I do need to replace the router. I wish there were some way to troubleshoot the problem first though: I'd hate to go through the hassle and expense of replacing my router, only to find out the problem lay elsewhere...
posted by ManInSuit at 1:43 PM on February 22

Argh. Need more reading comp. Yes, inSSIDer is for wireless.

Nthing trying a reboot, though. Turning off (we actually unplug the thing) and waiting a few minutes before turning back on.
posted by blurker at 1:47 PM on February 22

I have at times tried what a few people seem to be suggesting here: Unplug the router. Wait a couple minutes. Turn it back on. That sometimes works. Sometimes, I think, I have to wait another five minutes after the router comes on before the net connection works again (I'm pretty sure this is right. I'll try to start being more systematic in tracking the problem).

It's very hard for me to tell from this whether power cycling the router actually had any effect since, as I mentioned, the problem also fixes itself after a few minutes, even if I don't do anything.
posted by ManInSuit at 1:54 PM on February 22

How old is your router? I had a very similar problem with mine a couple of months ago--connection would bounce up and down intermittently. Sometimes restarting it would help, sometimes not. Often it would fix itself.

It finally went away when they came out and replaced the router. The technician said he'd done that three or four times that day.

If you can borrow some different hardware and swap it out, that should prove it definitively but I wager it's aging equipment.
posted by fifteen schnitzengruben is my limit at 2:01 PM on February 22

Another vote for dying hardware, likely caused my overheating - not difficult for those little boxes. When mine was going, I'd lose internet daily but the wireless connection would still be live.

When I bought my latest router I spent a little more ($70 or so), and have had zero problems for a year.
posted by shinynewnick at 2:07 PM on February 22

Does it happen when the wind blows? It could be some intermittent issue with the connection between your house and the utility pole.
posted by XMLicious at 4:04 PM on February 22

Most everything in your description sounds like a temporary DNS outage. I'd test that hypothesis by changing the DNS on one machine from unblock-us.com to a more conventional DNS like Google's 8.8.8.8 or maybe your ISP. Next time everything fails check the one machine with an alternate DNS configuration.

I'm less inclined to believe your router hardware is failing; the usual symptom of that is losing a whole connection (often wifi), not just DNS. It could be a software problem, although Tomato is pretty reliable. I'm assuming you're set up Tomato to not tell all your computers to use it, the router, as the DNS server? If you are try disabling that, although in my experience the Tomato DNS proxy is good.

If you do decide to replace the hardware, the ASUS RT-N16 is the modern replacement for the WRT54GL. It runs Tomato just fine.
posted by Nelson at 4:18 PM on February 22 [1 favorite]

Nelson - Yeah, I thought the problem might be related to the unblock-us DNS. From my own really uninformed impressions - it looked more DNS-y than hardware-y. But a couple of things make that diagnosis seem unlikely:

1) As I mentioned- when the problem is occuring: if I use the ping tool build into the router's admin page, the router has trouble pinging machines specified by hostnane, but not by IP. But the router is not configured to use the ublock-us DNS (it gets its DNS from our ISP). So presumably if the problem were just related to the DNS at unblockus, the router's own ping tool would be unaffected.

2) I'm pretty sure that we had this problem on all our machines, even when not all of them were using ublock-us. (Again- I'm like 95% sure of this. Again - I wish I kept better records of my troubleshooting...)

(I'm a little hesitant to try de-configuring unblock-us, because all the machines now use it regularly, and the inconvenience of a week without unblock-us is arguably greater than that of the occasional service interruptions...)

Thanks for the tip re: RT-N16!
posted by ManInSuit at 4:44 PM on February 22

Oh! I think I figured something out, which seems like it might be related to the problem.... The problem first started happening when mi ISP replaced my old modem - A cellpipe 7130 which was a little flakey, with a new one - a sagemcom 2864. Both modems are modem/router combinations, but I use them with my existing router. So... On the old modem, I had configured things so that the modem would just work as a modem, and disabled all the router functions. (I think this is called "bridge mode".). When I got the new modem, I just plugged it in, and everything seemed to work, and I never thought about it. But I never switched it over to bridge mode! This seems like it might be the cause of the problem, for a couple of reasons:

1) The problem only started when I swapped in the new modem
2) From my incredibly limited sense of how networks work, the problem I am having seems like the sort of problem you might have when two routers are battling to do the same thing on the same networks- not a periodic loss of connection (as I'd expect from, say, an overheating router) but periodic weirdness where some things work and some things don't. I don't know if I'm on the right track here.

I noticed when, following the advice of folks here, I went out and got a new router and started setting it all up. I may keep the new router anyhow (it's an ASUS RT-N66U, a step up from the old one I had) and so if the problem does go away, I may never know if the problem was the old router or the misconfigured modem...
posted by ManInSuit at 6:29 PM on February 24

Ah! Spoke too soon. Having completed the steps above, the problem seems to continue, even with a fancy brand new router installed, and DHCP and wifi turned off on the modem. I continue to get the same weird problem - a periodic inability to load web pages on all machines, even though I can still ping remote sites...
posted by ManInSuit at 10:11 PM on February 26

unblock-us is being very glitchy right now, if you're still having the problem.
posted by scruss at 10:29 AM on March 17

Thanks, scruss. But since posting the Q, I now have a machine on my network that is not using unblock-us, and it experiences the same problem.

I did recently get possible diagnosis from my ISP (teksavvy): It seems that the problem I describe is a know problem with my modem/service called "sync no surf". Apparently they can fix it by changing a setting at their end. I hope it works! I mention it here in case anyone finds this page trying to solve a similar problem...
posted by ManInSuit at 3:00 PM on March 31