Cisco Systems UBR10012 manual Cache Parity Errors, Sigerror, Sigreload

Chapter 4 Troubleshooting Line Cards

General Information for Troubleshooting Line Card Crashes

Table 4-2 SIG Value Types

SIG Value	SIG Name	Error Reason
22	SIGERROR	Fatal hardware error
23	SIGRELOAD	Software-forced crash

Step 8 The vast majority of line card crashes are either Cache Parity Exception (SIG type=20), Bus Error Exception (SIG type=10), and Software-forced Crashes (SIG type=23). Use the following sections to further troubleshoot these problems:

•Cache Parity Errors, page 4-4

•Bus Errors, page 4-5

•Software-Forced Crashes, page 4-6

If the line card crashed for some other reason, capture the output of the show tech-supportcommand. Registered Cisco.com users can decode the output of this command by using the Output Interpreter tool, which is at the following URL:

https://www.cisco.com/cgi-bin/Support/OutputInterpreter/home.pl

Step 9 If you cannot resolve the problem using the information from the Output Interpreter, collect the following information and contact Cisco TAC:

•All relevant information about the problem that you have available, including any troubleshooting you have performed.

•Any console output that was generated at the time of the problem.

•Output of the show tech-supportcommand.

•Output of the show log command (or the log that was captured by your SYSLOG server, if available).

For information on contacting TAC and opening a case, see the “Obtaining Technical Assistance” section on page x.

Cache Parity Errors

A cache parity error (SIG type is 20) means that one or more bits at a memory location were unexpectedly changed after they were originally written. This error could indicate a potential problem with the Dynamic Random Access Memory (DRAM) that is onboard the line card.

Parity errors are not expected during normal operations and could force the line card to crash or reload. These memory errors can be categorized in two different ways:

•Soft parity errors occur when an energy level within the DRAM memory changes a bit from a one to a zero, or a zero to a one. Soft errors are rare and are most often the result of normal background radiation. When the CPU detects a soft parity error, it attempts to recover by restarting the affected subsystem, if possible. If the error is in a portion of memory that is not recoverable, it could cause the system to crash. Although soft parity errors can cause a system crash, you do not need to swap the board or any of the components, because the problem is not defective hardware.

Cisco uBR10012 Universal Broadband Router Troubleshooting Guide

4-4 OL-1237-01