Cisco 2960 tftp problem

Answered Question
Feb 19th, 2009


We have several Cisco 2960 switches and we have a problem after upgrading IOS.

We had : c2960-lanbasek9-mz.122-46.SE

No problem at all transfering IOS files with tftp

Now : c2960-lanbasek9-mz.122-46.SE

With this IOS, every transfer crash. Only a few bytes are transfering then everything stop. On the TFTP server we see the connection disrapering (so the switch stop downloading).

If I restart the switch in the original IOS, everything works.

Other question, can you tell me what's the purpose of the file named "multiple-fs" in the flash.

Thank you

I have this problem too.
0 votes
Correct Answer by Tshi M about 7 years 8 months ago

is the same tftp server used in the lab as in the prod environment?

I would try to use a different tftp server (i.e. like your workstation) just to rule out the server.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Tshi M Thu, 02/19/2009 - 07:09

Hi Jean,

Both of the posted IOS are the same. Also by every transfer crash, do you mean the switch actually crashes or the transfer times out?


jphilippe.halle Thu, 02/19/2009 - 07:17

Sorry my mistake

Original IOS :c2960-lanbase-mz.122-35.SE5

Now : c2960-lanbasek9-mz.122-46.SE

The transfer just stop and nothing happens (I waited several minutes without seeing a time out).

Tshi M Thu, 02/19/2009 - 07:21

have you checked to see if you have enough space to uplaod a new flash?

dir flash:

I ran into a similar problem last week with one of my 3560.

I also had to increase the line vty exec-timeout but that might not be your case.

jphilippe.halle Thu, 02/19/2009 - 07:40

hi again,

I checked and I have enough space (more than the double).

To add details, we upgaded the IOS because we needed crypto features.

I don't understand. I took a another 2960 spare switch and put the same IOS' and I don't have the problem...

Tshi M Thu, 02/19/2009 - 07:44

Interesting fact. I'd open a TAC case if I were you. The problem seems to be isolated to that specific switch.


jphilippe.halle Thu, 02/19/2009 - 08:39


I think I have found the problem. The server running the tftp server is a windows with NIC Teaming and I think it's the problem. I removed the teaming and now the download from the switch works almost (now instead of crashing right at the begining it's crashing at ramdom % and I see CRC errors on the NIC of the server).

I'll continue investigating

Tshi M Thu, 02/19/2009 - 08:40

I will check the settings on the switchport where the server is connected. You might have a duplex mismatch.


jphilippe.halle Thu, 02/19/2009 - 09:10

Seems to be ok for speed/duplex...

If I tranfer the IOS via tftp (UDP) it's always crashing during the download. But for some reason it works perfectly over HTTP (TCP).

Am I correct to assume that we have bad packets on the network and it works over HTTP because TCP retransfer failed packets ?

Tshi M Thu, 02/19/2009 - 09:12

I am just confused cause it works on the spare switch.

jphilippe.halle Thu, 02/19/2009 - 09:20

the spare switch wasn't on the same network. It was in our lab

and it didn't work on all the switches in the same network

Tshi M Thu, 02/19/2009 - 09:25

Can you upload from the switches to the tftp server without problem?

I would also span the server port just to see what is happening when you do your transfer.

jphilippe.halle Thu, 02/19/2009 - 09:33

yes, all tranfer from switches to the tftp server works.

Only problem is from tftp to switches.

And only with the new IOS...

Correct Answer
Tshi M Thu, 02/19/2009 - 10:46

is the same tftp server used in the lab as in the prod environment?

I would try to use a different tftp server (i.e. like your workstation) just to rule out the server.

jphilippe.halle Fri, 02/20/2009 - 06:14

no it wasn't the same tftp server.

I pretty sure the problem is the computer we are using as tftp server.

I will test that later

Thanks for your help


This Discussion