re-examine the ables you're using to connect from switch to switch. Take out all cross-over cables (used to connect the hubs) and replace them with straight cables.
Main Topics
Browse All TopicsI've had a puzzler here. Wanted to upgrade from some old 10-BaseT hubs to 10/100 switches--I'll just describe one of those upgrades here. The office has 50+ IP Addresses active (PCs, Linux Boxes, Printers, etc.) all going through two 32-port switches. This works fine. in one office, we have 4 printers and a workstation connected to a 10-BaseT hub. I swapped it out with a 10/100 switch. At first it all worked fine. I went home, came back the next morning, and I'm getting all kinds of complaints about latency (it's taking 5 minutes to print a 2.5 MB excel document). When I "switched" back to a 10-BaseT hub, it all worked like a charm.
I've seen this same thing in other offices--5-port D-Link switches, 8-port Gigafast switches, ...doesn't matter--they all are flaky and slow.
What could be going on? Am I missing something about how switches work?
Dean
This Question has been solved and asker verified All Experts Exchange premium technology solutions are available to subscription members.
Experts Exchange has been collecting answers to technology questions since 1996…3 million and counting! If you have a question, chances are we already have your answer.
If you can't find the exact answer you're looking for, ask our exclusive community of 50,000 experts. You’ll get a personalized answer from a trusted professional.
Thousands of free tech tips, tricks, how-to’s and tutorials are available in our peer reviewed articles section. See for yourself how smart our experts are, no login required.
Access the answers to your technology questions today.
30-day free trial. Register in 60 seconds.
Members of the expert community talk about why the experience at Experts Exchange is different than what you will find anywhere else.

Try it out and discover for yourself.
30-day free trial. Register in 60 seconds.
Join the community of experts here and help other tech pros by answering question in your area of expertise. You can earn FREE access to all Experts Exchange's premium features and resources.
I would bet that you have 2 cables connecting between the 2 hubs or switches somewhere.
Do this :
Identify the main switch (usually the biggest switch) have each hub (or smaller switch) have a Cat5 to this main switch. Connect it all together. go to each hub (or smaller switch) and put a ping -t to the main switch (or a PC on that main switch) Unplug the uplink from the switch to the hub, if it times out you have a single connection, if it continues to ping, you have more than one connection back to the main switch. Do this for each hub. Also you absolutely do NOT want the hubs connected together.
If you want to have 2 main switches (with a connection in between the 2 main switches), connect each hub (or smaller switch) to either one switch or the other, not both.
A couple more thoughts:
--------------------------
Unplug the uplink from the switch to the hub, if it times out you have a single connection, if it continues to ping,
--------------------------
The reverse could also be true. If you cannot ping (or have very high pings) and you unplug the cable running back to the main switch and the pings go back to normal, that would indicate 2 links (one to the switch and one to somewhere else on that hub)
If you have high pings (abnormal) and when you unplug the link back to the switch, and it times out, that would indicate that one of the other hubs connected to the switch has a loop. In this case, leave this hub disconnected and go to the next hub. Same? leave it disconnected and go to the next hub, etc until you find the one causing the issue.
Now, specific to you, since this one hub (when changing it to a switch) is causing the issue, I would think that A. it has more than one connection to the main switch (via 2 cables connecting to the main switch or via 1 cable connecting it to the switch and another to another hub that also has a cable back to the main switch), or B. it is connected to itself somehow.
for problem A use the method above.
for problem B use the method below.
To find out if it is connected to itself, disconnect each connection and make sure only that link light goes out, if another link light goes out on the same switch/hub, you have found your loop.
Thanks for the feedback so far. A couple of additional points of information. We took that switch out of its location connected to the 4 printers and desktop PC, and put it somewhere else connected to two other PCs (new Dell desktops) and it seems to work fine, though we have noticed that it runs very hot. Over several hours this could be an issue. We're checking this now. But even so, I've seen this same problem with other switches in other locations in this office. These are all autosensing switches and no cross-over cables were used. (The hub of course has an uplink port). Nothing loops back. The cable comes from the wall to the switch (plugged into port 1) and the 4 printers and desktop are the end of the line here. T
I don't believe the cables are the problem per se because I can swap out the switch for a hub using the exact same cabling and it works like a champ. In another location, I replaced a cheap 8-port Gigafast switch with a hub and it worked fine (same cables), and then swapped it with another switch (a 16-port D-Link switch) and it worked fine. If the cables are a problem, could it be that some switches are more sensitive to old cables? Our office was wired 10+ years ago, and every office has wall porThanksts connected to the 2 32-port switches in our network closet.
Sometimes its a good idea to statically define the speed and duplex on the switch for the devices connecting to each port. The negotiation for autosense is often unreliable.
If you are concerned about Spanning tree issues, you should research how to set the root bridge manually as well. All of the switches out of the box will have the same priority (if they are the same model) As you add switches to the same layer 2 domain, you'll eventually have problems with STP.
A switch's function is to ROUTE traffic BETWEEN different IP domains, whereas a router will assume they are all of the same class C domain. Regardless of the "running hot", I would suspect your switch is trying to split traffic to two different subnets, and the printers are on a different subnet than most of the computers. Therefore it is trying to route that traffic to a non-existant, or non-vald subnet. Why not just use a router if your computers and printers are all on the same subnet. What else are you trying to do?
had the same probelm in my work, for some reason the Switch was running the port at Half Duplex even though the Server and switch were on auto neg. there were problems saving files to the server and printing.
I forced the switch to run the port at full and problem solved.
if that doesnt work to a port mirror and put a port analyzer on there, see what's going on (long acknowledgements etc).
you can get one here
http://www.ethereal.com/
http://www.solarwinds.net/
Exactly what kind of swtiches are you putting in? If everything works great on your old 10baseT hub, are all your hosts configured for Autonegotiate? If so, then all the ports on your switch also need to be Autonegotiate. On the other hand, if you want to hard set your network to 100 Full Duplex, then all the hosts need to be 100 Full Duplex as well.
Can you get into the switch and examine the Ethernet port statistics? What kind of errors are you seeing? If FCS & Alignment Errors, first thing is to make sure your duplex is set correctly on both sides. You cannot have only one side Auto unless the other side is SET to HALF.
Hmmm, 10+ years is a long time. Did you replace a 10 Mbps hub with a 100 MBps switch or something like that? If the cable is not CAT5 or better maybe it is having problems negotiating 100Mbps.
If the switch runs hot, it could definitely cause issues when giving it a load.
I don't know what " A switch's function is to ROUTE traffic BETWEEN different IP domains," means, but I know that a switch's function is to Switch conversations to different MAC addresses. Your simple switch does not even care about IP unless you get into higher end switches, and that does not seem to be the case here.
The only topology difference between a simple hub and a simple switch is that a simple hub has only one collision domain, and a simple switch has a collision domain for each port. Of course, pay more money, get more features like VLANS, 802.1q, routing built in, etc.
SO..... Do this:
Hook up the switch like it was before (when causing issues)
Put a continuous ping from one workstation on the switch to another workstation on the switch.
Put a continuous ping from one workstation on the switch to another workstation somewhere else in the building (so the ping has to go thru your switch backbone.)
When the pings stop or get really laggy, disconnect the uplink to the back bone.
If your pings to the other workstation on the same switch normalizes (drops back down to an acceptable ping) you know your problem is with the wiring.
If your pings to the other workstation on the same switch stays high (or intermittent) you know that this switch is bad.
Business Accounts
Answer for Membership
by: pclark6127Posted on 2005-06-01 at 13:55:56ID: 14125859
Check for duplex mismatchs on the switch. Also, if they are Cisco check to see if port-fast is enabled and if so disable it. Also check cabling to make sure it is good. One other thing to check is make sure you do not have a loop.