I just moved my computers over to a new switch. Everything works fine, but given enough time, usually about a day, any number of my computers on the switch will not be able to communicate via one another. I can't ping them from anywhere, and they cant ping anything obviously.
Lets say I get a call saying the computer A plugged into port 10 on the switch isnt functioning. The link says I am physically connected and the software on the computer and the switch says it is enabled but I cant ping Computer A. When I move that computer A from port 10 to say port 11 on the switch I am then able to ping computer A and everyone is happy. Furthermore when I plug a known functioning computer B into port 10 I can no longer ping the computer B when it is plugged into port 10.
Resetting the the computers doesn't fix the problem. When I reset (power cycle) the switch 29xx or 35xx (Im not sure which one) everything works fine... for a while. The computers I am experiencing problems with are static IP and the ports and nic speed/duplex were set manually matched to 100 full on both ends.
The reason I moved to a new switch was that every so often ( no more than 3 times a week) my actual physical connection would disconnect/reconnect disconnect/reconnect like a power cycle almost randomly. The 'power cycles' never caused continual communication problems like the current setup.
I'm not really looking for a detailed explanation on my current setup issues, just an affirmation that my logic in thinking something is wrong(hardware/software config) with the switch is correct.
Does anyone think it could be a computer or computers on my switch causing this problem and not the switch?
Any ideas what could cause a physical disconnect/reconnect disconnect/reconnect?
Thankyou very much.
You may able to reset the port by using "shut" & "no shut". You have to check the log of the switch to determine any error or warning message to say the computer A connected port problem.
Moreover, did you check is there any virus in the PC that create the network problem ? I suggest to make a virus scan in all PC to ensure there is no software issue then ensure the NIC is error-free too.
There is no way to troubleshoot it step-by-step.
Hope this helps.
It sounds like a very strange problem indeed.
Are you able to access the switch remotely?
If so, do you have a configuration of the particular port in question?
Did you find that the last problem was definitely a problem with the switch? I.e. does the old switch still do it if you have a few cables plugged into a computer at your desk?
To get more detailed information, maybe next time it happens you could get a bit more information:
1. on the switch do a "show inter fastethernetx/y" to see if line protocol and pysical is up
2. issue a "show mac-address-table" and see if there are mac addresses for that port in its table
3. type "arp -a" on your PC and see if there's an arp entry for the gateway
By your last statement are you asking what could cause your issue with the last switch, or the current one you're having problems with?
Yes, the switch can be accessed remotely.
There were a couple different networks plugged into the old switch and everything connected to the switch would 'power cycle' as witnessed in event logs on all local computers.
Unfortunately, its not my switch and I dont have access to it. I do know that when port 10 was down, through some gui on the switch port 10 was disabled/enabled and it still didnt work. I removed my default gateway.
I was asking what would cause the issue with the last switch?
in addition to the other posts, I had a similar issue a while ago. After doing a lot of research without any result, I had to eventually replace the switch. The 2900/3500XL switches are End-of-life/End-of-Sale, so you might simply have a hardware problem. If possible, have the administrator of the switch try and connect a device directly to whatever is connected to the switch (most likely a router), and check if the problem persists.
Sounds like link flapping to me - you need to have the switch administrator check the errors on the switch port - runts would be indicative of a faulty NIC in your computer - have you tried changing the NIC and plugging the computer back into the original switch port.
Switch administrator would know more - you need to speak to them!
Hope this helps
I'm not sure what it was, but we plugged everything back into the old switch (the one that will power cycle 2 times within about a minute around 6-10 times a month) and have not had any communication problems minus the infrequent 'power cycle' So I definitely think it was clearly the switch, whether or not it was a hardware failure or configuration issue is still undecided. But the network admin was sending the switch back to Cisco I believe.
I don't think it would be a faulty NIC. I had about 20 pcs plugged into the switch and for no rhyme or reason any or all of them would lose communication ability although they would still be physically connected. So you think I have 20 bad NICs? Is it normal/possible for one bad NIC to bring down an entire switch? I would think a bad NIC would just affect the port it was currently plugged into.
If there are 20 NICs w/ the same problem of the switch. If the 20 NICs are the same brand and model, tt may be the compatiblity issue more than failure issue. If thw 20 NICs are different brands, I believe it is the failure switch.