3550 switch loses routing table and displays memory error

[m.tapping]

3550 switch running IOS (12.1(19)EA1) loses routing table entries and locks-up. Plugging a console cable into it receives the following message "out of memory, please try again later". This is also mirrored on other directly connected (GBIC) 3550 devices. A power cycle clears the problem.

I have checked some IOS info and not found anything of value to-date. Has anybody seen anything like this before? is this a known issue?

I'd appreciate any help.

[Tsasbrink]

How large is your regular routing table, what routing protocol are you running ? Do you have more of these switches with the same software without these problems ?

[aashish.c]

Hi,

It is CDP memory leak bug in 12.1(19)EA1. I dont remember the Bug no. but u need to upgrade the IOS to 12.1(19)EA1c or latest 12.1(22)Ea1.

In this bug CDP process eats up the memory and switch keeps giving "low in memory, please try again" error.You can also tuen off the CDP by :

switch(config)#no cdp run

after this command reload the switch so that switch releases the captured memory and you will not get this error message.

Regards

aashish C

[Kevin Dorrell]

CSCdv56078 ?

[pflunkert · Accepted solution]

Hi,

CSCec89172

Release Notes

The Catalyst switch runs out of memory due to a CDP memory leak.

This only affects Cisco IOS Release 12.1(19)EA1. The following devices are

supported in Cisco IOS Release 12.1(19)EA1: Catalyst 2940, 2950, 2950-LRE,

2955, 2970, 3550, and 3750.

This problem does not affect any other Cisco IOS Release.

Workaround:

1) Downgrade to 12.1(14)EA1 or earlier release.

2) Disable CDP using the global configuration command no cdp run.

3) 12.1(19)EA1a and future releases have the fix for this problem.

Regards

Peter

[m.tapping]

That's very useful, thanks. I've reviewed the software notes and we are planning to upgrade in the next few days.

Regards

Marc.T