Network Connectivity Consistently Drops

Posted on 2013-06-04
Last Modified: 2013-06-07

For the last couple of days we have been experiencing more than the usual network lag.
I blocked some of the more bandwidth intensive websites, and this did not help. I then noticed that with my initial troubleshooting that a ping –t to my router (which is also my proxy server appliance), as well as my servers (domain controllers, print servers, etc) would show a periodic timeout. This would also be evident when I had remote desktop sessions to my servers, the session would have to reconnect itself.

All of my servers are connected into the same switch in our server room, which connect into our main switch which resides in data closet.

Pinging from anywhere in the building yields the same results. I have not found anything in the server logs indicating an issue. I had our proxy vendor run internal tests on our proxy/router, and it came back clean. I have replaced the switch in our server room & data closet. Replaced cables, used the alternate data drop for the connection between our servers and data closet, and the result remains the same.

Also, we are running Server 2008 R2 on all servers, clients are running a mixture of XP, vista & windows 7.
Question by:ibewlocal98

Expert Comment

ID: 39219815
Has anything recently changed on the network? Do you have Spanning Tree enabled to ensure that there is not a loop in the network causing a broadcast storm and the issues you are seeing?

Does this seem to be related to a certain time of day each day or is at random?

Accepted Solution

SmallBall earned 300 total points
ID: 39219833
Sniff out your network for excesive/illegal/bogus  traffic , and chceck switch logs for loops (spanning tree should be on if u have +1rep for bigpapagotti ;)), if it fails - connect your network on short  (1 client + few servers) and sistematicly plug in rest of your nodes while chcecking dellays.

Author Comment

ID: 39219864
thanks for the replies, and let me mention I am not where i need to be with the networking skills.

When the problem first presented itself, nothing on our network had changed. It was after I switched out cables and switches.

Also, no spanning tree and our infrastructure is old and inexpensive. No logs from switches, most are generic trendnets.
Efficient way to get backups off site to Azure

This user guide provides instructions on how to deploy and configure both a StoneFly Scale Out NAS Enterprise Cloud Drive virtual machine and Veeam Cloud Connect in the Microsoft Azure Cloud.


Assisted Solution

BigPapaGotti earned 100 total points
ID: 39219915
Can you explain what you mean by "When the problem first presented itself, nothing on our network had changed. It was after I switches out cables and switches"

I'm assuming that you mean the network slowness as the the problem that first presented itself. But then I am confused as to what you mean by "it was after I switched out cables and switches" Do you mean after doing this the network really bogged down and became unresponsive?

To check for a loop please ensure there are not multiple cables connected between the two switches. This means that Switch 1 must have one connection to Switch 2 and Switch 2 must have one connection to Switch 3. Switch 3 CANNOT directly connect back to Switch 1 as this would cause a loop. Think of the connections as being daisy-chained together. I would ensure this is the cause in your network topology.

Author Comment

ID: 39219967
What I mean by this is I did not make any changes until there was a problem. After I noticed the pings dropping, I changed out switches and cables.

Author Comment

ID: 39219972
Also, from our server room there is only one connection to the data closet.
LVL 24

Assisted Solution

smckeown777 earned 100 total points
ID: 39220138
Working on the same path as @SmallBall I'd do the following

Connect server and 1 client to the main switch - from that client run a continous ping to the server like

ping -t <server ip address>

Now start connecting each of the other clients back into the network - how many clients have you? Assuming you've a 24-port switch shouldn't take that long...

Plug each client into the switch - leave 30-60 seconds between each client you reconnect - watch the ping on client1(i.e. if you've a laptop use that so its close to the switch you are working on)

Once you connect the rogue client it should cause the ping to start dropping...

Only reason this may not work is if these 'ping timeouts' that are currently happening are very random - can you tell us how often you've seen it happening when running ping tests? I mean is it every 10 pings? 20? 30? The longer it is the harder it will be using this method to pinpoint if its a client that is causing this issue...

Expert Comment

ID: 39221913
I had an issue years ago with a loop on a basic layer 2 switched network infrastructure.

All the floorsockets were patched into the distribution switches and someone had patched one floor socket into the one next to it to keep it tidy... sigh...

depending on how big your office is i would give this a check - look for floor sockets patched into the one next to it.

took me 2 weeks to uncover this nugget of info and after that the problems dissapeared.

all the best


Author Comment

ID: 39221944
I am in 12 story building. We acquired more floors as time went on. Didnt have all 12 floors initially, so as we added we basically daisy chained floor to floor, then into main closet. I am in process of trying what small ball suggested, disconnecting everything except laptop and connection from server room. Will post results

Expert Comment

ID: 39222356
Sounds like you created a loop by your statement of "daisy chained floor-to-floor then into main closet" I could be wrong though

Author Closing Comment

ID: 39228829
thanks for all the help. Unfortunately in our setup, everything is basically daisy chained floor to floor, as we acquired each floor at different times.

So I disconnected everything from main switch except our router and connection from switch where servers reside.

I then systematically swapped out the switch that controls each floor. Found problem on last floor :) Issue was actually being caused by the switch in our computer lab on bottom floor. Disconnected this, replaced switch in server room to be sure, and all is well.

thank you for all your help. And I will take this as a lesson in why you shouldnt daisy-chain your network.

Featured Post

Netscaler Common Configuration How To guides

If you use NetScaler you will want to see these guides. The NetScaler How To Guides show administrators how to get NetScaler up and configured by providing instructions for common scenarios and some not so common ones.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

This article is a how to to configure a UCS Ethernet-uplink portchannel via the console. It is easy to do and can be done quite quickly. In certain versions of the UCS manager the portchannel has issues coming up and this is a workaround. I am…
This is the first one of a series of articles I’ll be writing to address technical issues that are always referred to as network problems. The network boundaries have changed, therefore having an understanding of how each piece in the network  puzzl…
Windows 10 is mostly good. However the one thing that annoys me is how many clicks you have to do to dial a VPN connection. You have to go to settings from the start menu, (2 clicks), Network and Internet (1 click), Click VPN (another click) then fi…
Finds all prime numbers in a range requested and places them in a public primes() array. I've demostrated a template size of 30 (2 * 3 * 5) but larger templates can be built such 210  (2 * 3 * 5 * 7) or 2310  (2 * 3 * 5 * 7 * 11). The larger templa…

776 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question