Field Notice: ONS 15454 ML Series Card Memory Allocation Issue

Revised February 11, 2005

January 31, 2005

---

Products Affected

Products Affected
15454 - 15454-ML-100T-12
15454 - 15454-ML1000-2

Problem Description

When bridging over a resilient packet ring (RPR) shared packet ring (SPR) interface or a port-channel interface, MAC address aging may lead, over time, to a memory allocation issue.

Each time a MAC address is placed in the switching table, and later removed due to inactivity, the control plane loses track of a small amount of memory which is not reclaimed.

In time, the control plane will be depleted of all memory, and will stop functioning. The length of time for full memory depletion will vary based on the activity of the network and number of host devices directly connected via Layer 2 switches. The time period that has been seen varies from weeks to many months. At that point, a number of error messages, including %SYS-2-MALLOCFAIL , will be seen on the console, and login attempts to the card will fail.

In most cases, traffic through the card will continue because the data plane will continue to function. However, RPR protection, spanning-tree protection, and port state changes (link up/down) will not function, and this may cause loss of traffic, and the CONTBUS-IO alarm may be raised to CTC/CTM .

Note: The RPR/STP mechanism, on a card experiencing this issue, will not switch and any fiber/span failures will cause traffic loss on all flows passing through this card. To prevent an outage, it is highly recommended that customers implement the workaround listed below or upgrade to software release 4.1.8, 4.6.4 or 5.x.

If routers running Layer 3 routing protocols are the only devices connected to the ports of the ML-series network, this problem may not occur because there will be no inactive MAC addresses to remove.

Background

If there is a port-channel or SPR interface which L2 is learning MAC addresses from, this problem will occur. First seen in software release 4.1, this issue has been fixed in minor releases 4.1.8 and 4.6.4 and in all releases beginning with 5.0. There is also a short-term workaround to this solution in the Workaround/Solution section below.

Problem Symptoms

Telnet will not connect to the card. Login via console fails with memory allocation error messages. Console output shows periodic error messages including %SYS-2-MALLOCFAIL messages. CONTBUS-IO may be raised by CTC/CTM . A card reset is needed to restore the management connection.

Workaround/Solution

This issue is fixed in minor releases 4.1.8 and 4.6.4, and in all releases beginning with 5.0.

The following workaround may be used in the interim, prior to upgrading to one of the software releases which corrects the issue, as listed above.

To reduce memory loss, increase the MAC aging time so that short periods of inactivity will not trigger MAC aging. Using the maximum aging time of 1000000 seconds, about 11 days, should prevent nearly all memory loss, ensuring that the issue does not occur.

Set aging time to the maximum with the following configuration command for each bridge group X :

bridge X aging-time 1000000 
bridge X limit dynamic entries 10000 
exit 
copy running-config startup-config

Once problem symptoms occur, a card reset is required. If reset via CTC fails, it may be required to physically remove and re-seat the ML card to initiate the reset.

DDTS

To follow the bug ID link below and see detailed bug information, you must be a registered user and you must be logged in.

DDTS	Description
CSCeg11560 (registered customers only)	15454 ML memory allocation issue, console/telnet fails. Possible CONTBUS-IO alarm.

For More Information

If you require further assistance, or if you have any further questions regarding this field notice, please contact the Cisco Systems Technical Assistance Center (TAC) by one of the following methods:

• Open a service request on Cisco.com
• By email
• By telephone

Receive Email Notification For New Field Notices

Product Alert Tool - Set up a profile to receive email updates about reliability, safety, network security, and end-of-sale issues for the Cisco products you specify.

---