On high-end SRX devices, an SPC crash may result in a core dump being generated without an alarm and a "Machine Check" reason string to be reported in the nvram command output of the services processing card (SPC).

This article lists the causes for the SPC to crash with the "Machine Check Error" and explains what must be done to resolve the issue.

 

SPC crashes and a core dump is generated. On high-end SRX devices, a core dump caused by an SPC crash might not generate an alarm.

The crash might be caused by the machine and the reason for the crash can be determined from the output of the nvram command executed on the SPC shown in the RSI output.

=== SPC4, node0.fpc4> show nvram

System Exception: Vector/Code 0x00200, Signal 11
Event occurred at: Jan  7 14:06:15.12697

Juniper Embedded Microkernel Version 12.3X48-D30.7
Built by builder on 2016-04-28 21:31:05 UTC
Copyright (C) 1998-2016, Juniper Networks, Inc.
All rights reserved.
Reason string: "Machine Check"
Context: Thread (XM Chip Statistics)
Machine Check address reg 0x418a3f60, Syndrome 0x20000000
  Data cache push parity error <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< 

 

The SPCs may crash/core-dump with the reason "Machine Check" as seen in the output shown above, due to the following reasons:

• Machine check input pin

• Instruction cache parity error

• Data cache push parity error

• Data cache parity error

• Bus instruction address error

• Bus Read address error

• Bus Write address error

• Bus Instruction data bus error

• Bus Read data bus error

• Bus Write bus error

• Bus Instruction parity error

• Bus Read parity error

 

All the above listed errors point to potential hardware issues and may require a replacement of the card. Contact Customer Care for the same.