Strange network problems at small biz

Appleby

New Member
Reaction score
3
Location
Texas
Ok time to call in the big guns for help. Here it is in a nutshell....

I've got a small business customer with 5 Dell pcs, all 1-3 years old, very nice clean machines. 1 older custom build pc used as a file server and 2 network printers. There is a laptop that is sometimes used wireless as well as several iPhones in the office. Cable ISP Motorola modem, Netgear wireless router and a 20-24 port switch. This is a very simple setup and it's worked fine for many years. Over the last 3-4 years I've replaced 3 different routers that seem to just die. Very strange but we've all seen these $40-$60 routers do that. They've just been through several.

I get called out a few weeks ago because no one can get online, wired or wireless. I go out, test the router and I get no internet access with my laptop, wired or wireless. Power cycle the whole deal, nothing. There is a connection when directly connected to cable modem. I plug in a new router and boom everyone is back in business. Next day they call and say that half the computers are working and half aren't. I go out and all have pulled local IPs but half can't ping out. Power cycle router all back online. Next day, same thing when they come in at 8 am. Some are online, others not. I go back out and start to wonder if they have a switch problem. I didn't have a large enough switch with me, so I connected two new 8 port switches and left them working.

Same thing keeps happening nearly every day. I consider a possible power issue and they have a very old power strip where the modem, router, switches are plugged in, so I replace the strip to a new surge protector. Again, everyone works until late in the afternoon when most of them go offline. They power cycle modem and router, all is well again. Then it stops happening during the day, but when they come in at 8 am the computers can't surf. Power cycle and they usually work fine all day.

I assign half the pc's a static IP just for giggles and I change the NIC in one of the computers that seems to be the first offline every time. No change at all. I pull the 2-3 of the RJ-45 wall plates to check the wiring and it's perfect. Actually the best punch downs I've ever seen and NO spare wire stripped back too far. I'm impressed. I go ahead and use my handheld network cable tester to see if there are misswires anywhere (couldn't check them all because they were very busy) but I checked 3-4 of the computers, all the way from the patch cable going into the back of the pc, all the way back to the switch and all is good. I replace the cables going from routers to switches just for grins. Same thing happens every day when then come in at 8am.

Guys I'm at a loss here. I can't figure out WHAT is causing this problem. The modem never loses a connection, so I feel like that's ruled out and much of the time, only some of the pc's lose their connection. They do have two computers that seem to be the same ones that are first to lose their connection when it happens during the day. But for the most part now, it's going out over night. And the computers never lose their local IP, they just can't ping out.

I'm pulling my hair out and they are being very understanding and the owner's husband is actually a big wig at the local cable ISP and he's come in and tested everything on their end and done some looking around and he is stumped too, which gives them some piece of mind that I'm not just crazy.lol

Is there something simple I'm missing? I've replaced all the networking hardware, except for all the NICs. I'm wondering if one NIC could be going out and causing the problem? Or maybe a printer?

HELP!lol :confused:
 
can they ping around the local LAN? You just said they cant ping out....
Depending on the lease schedule, they should be keeping their IP's, so thats not uncommon.... Might lower the lease time to an hour and see if that helps? Actually come to think of it i had something similar once. Turned out they were running out of IP's because of to many wireless devices sucking up all the addy's, and it would eventually kill everyone, so check the address range too...
 
Last edited:
Same thing happened to me at my own house a few months ago. Walk into my house, network goes down.

I figured out through the process of elimination that the problem was my new iPhone 4s. There is a bug in it's DHCP process that would cause the router WAN port to go offline every time I walked into the house. My router at the time was a Linksys WRVS4400N business router.

So I went out and got a Netgear router, thinking it would solve the problem. Got a fairly low-end one, but it had the same problem :( Returned it an got a very high-end Netgear N900 dual-band wireless router and problem solved!

Eric
 
Try a different router

You may just have a gooffy router. I would try a different one. I have had excellent luck with the Cheap Tenda Router, which I can pick up at my local microcenter for $14.99 US.
 
Appleby,

Based on your description, it looks like you might have a node that’s flooding the network. Sometime the network works and sometime just half. So I would say the Modem and router is probably ok. But but I would have the ISP run a loopback test to the modem and run some diagnostic on the router if possible. Here’s what I would do.

Option 1) I would temporary replace the 20-24 switch with a HUB and run Wireshark (24x7). If one of the workstation is trashing, you will see it on the capture.

Option 2) I would temporary replace the 20-24 switch with Cisco 3550 or 3560 Switch. Then enable loop guard, UDLD / storm control. If one of the Workstation/nodes decided to flood or cause problem, the switch will automatically turn it off and put it in error-disable mode. Then all you have to do is check which port is on err-disable and start troubleshooting from the port to the workstation

I bought my Cisco 3550-48 port at Ebay for $125. There are other Cisco Switch but I decided to go with 3500 because it’s also provides POE. Useful for troubleshooting VOIP. The disadvantage is it requires a lot of juice to power up. You can use 3500 series for option 1 but you need to configure the port for SPAN, or port mirroring.

Good Luck
 
Last edited:
Can you ping the default gateway?
What type of router? Have you upgraded the firmware?
Can you ping public IP instead of name?
Also, I would check to be sure your getting the IP from the router's DHCP.
 
Check for multiple DHCP servers handing out overlapping IP's, or incorrect IP/gateway settings...

Since the problem of half working/half not working started after replacing the router, that would be the first thing I would check... Perhaps the Motorola has a DHCP server in it that wasn't being used before and somehow got enabled, or was being used and now you've added another in the router.
 
Perhaps the Motorola has a DHCP server in it that wasn't being used before and somehow got enabled, or was being used and now you've added another in the router.

If the cable "modem" was a combo modem/router (like the SMC Gateway ahd Comcast typically uses for their business setups)...and the network had a router in between itself and that cable modem....DHCP does not pass through the common routers used in gateway mode. All my business cable internet setups have DHCP still enabled on the ISPs cable gateway..I have my own routers behind that..usually in static IP subnet setup. DHCP won't pass past the router.
 
I with synack2 here

Lets see, was a working network then dead. You troubleshot and replaced bad router. Working again but develops different symptom/problem. What has changed? I would go back and try another router. I have seen a lot of problems with cheap SOHO routers over the years and I would eliminate it before going much further since it was replaced and this new symptom developed.
 
To the OP......I'd look at a couple of different things.

As mentioned above...one of the PCs could be spewing out crap. Hard to believe...since they've had this issues for several years, you'd think that a PC which was spewing out junk on the network would have given more issues and been found by now (like hijacked by a rogue and turned into a spambot). But something that heavy can fill up the state table of typical home/SOHO grade routers and bring them grinding to a halt...requiring a reboot. For our business network clients....we usually start with true business grade routers...no "home routers" allowed, this is one of the reasons...they can maintain connectivity and performance under heavy loads.

I learned many years ago that network equipment can be sensitive to power...especially broadband routers. Little dips.."brown-outs"..greatly shorten their life. My standard setups for network equipment always have all core/edge network equipment plugged into APC battery backup units...just a little APC ES500 or so. They're quite inexpensive..and easily pay for themselves by allowing routers/modems/switches to last a long time. And the network runs nice and stable. Since staring to use a little APC at each client for their network edge equipment, I simply don't have the "another dead router" problems that I saw without them. A surge strip only helps against big spikes...but does nothing for brown outs..little dips.

Check location of the modem/router/switch....as for proximity near something giving off heavy EMI. I remember not long ago, going to a small car dealership/repo place/garage...their network stuff was located right next to the main electrical panel for the garage. They kept having network lock ups. Moved their network distro area...problem gone.
 
There is a laptop that is sometimes used wireless as well as several iPhones in the office. Cable ISP Motorola modem, Netgear wireless router and a 20-24 port switch.

Some are online, others not.

I have seen similar issues to what you're talking about with routers alone before, but with a switch behind the router I wouldn't expect to see it.

I've seen it before where the router only handles up to 8 connections at once. Wifi and Wired connections cumulatively up to 8. The DHCP on the router fouls up sometimes when you have more connections than this. If you're running some devices direct into the router, and others on the wifi, you might be surpassing this mark.

You might try running ONLY your wired connections for a few days (Just disable the wifi or change the PW so they don't have correct access) if the issue doesn't arise again, you would probably need to upgrade to a bigger router.

The other thing you could try that is not quite as "easy" but it works out better to maintain business and the current network. Try a separate wifi switch, and don't use the wifi on the router itself.

I know it seems odd, and strangely easy but its always a possibility.
 
This is one of the problems you are going to pull your hair out on.

Order I would check things in...

1.) Cable modem, is it using NAT or handing out public IP address to router? If it's NAT see if you can change that.

2.) Check for rogue DHCP servers, disable DHCP on your router and see if anyone gets IP addresses I've seen a company bring down an entire hotel this way thinking they were setting up a wireless booster into the office they were renting out of it.

3.) Drop in a managed switch as an earlier poster described, usually what I do is drop in a computer with 2 nics that reads the information going out to the router from the network but that is an amazing idea that I will have to implement.

4.) Check to make sure there isn't a cable that has been nawed on by rats or scraped across a sharp piece of metal that is bringing current into the network, this is a long shot but it's happened to me twice on seperate customers, use a cable tester if it suddenly shows linked cables (2 touches 4 on both ends) there you go.
 
Wow Great responses and things to test out guys. I appreciate the ton of help. I'm going to sit down and re-read all the replies and see what all I can reasonably try and see what info I can gain.

Thanks so much for the replies, I will report back when I know more, but I might not get over there until next week.

Thanks!
 
push

sorry for pushing this super old thread, but was this ever resolved? What caused the problem?
I would love to hear and find what happened.
 
I just caught wind of this thread believe it or not and Iam having a similar problem at a small office of about 8 computers and some wireless phones/laptops. I changed out the router for a linksys E900 and also down the hall I changed out a AP that simply would not configure. I then switched out the switch for a 10/100/1000 unmanaged 8 port switch. Whats wierd is that after a small remodel job to remove an access door the other office they dont have access to their AT&T DSL modem unless they call maintenance to unlock the door. I finally pressured them to get the maintenance guy to unlock it and found one of those motorola surfboard modems as their DSL modem. Its a pretty cheap thing.

They are going to upgrade to AT&T universe monday and Im going in to check everything again after they are done.

Is it really true about the off-the-shelf wireless routers kinda max out at 8 connections before trouble starts happening? Thats good to know. I might get that out of there.

HOpefully we can keep this thread going and learn the resolution to these types of problems.

Best Regards,

coffee
 
Back
Top