Article Detail

Description

This article provides information on how to replace any hardware component on an SRX1400/3400/3600 device that is configured in the chassis cluster.

Symptoms

How to replace a faulty SFB/RE on an SRX1400/3400/3600 in a cluster.
How to replace a faulty NPC card on an SRX1400/3400/3600 in a cluster.
How to replace a faulty SPC card on an SRX1400/3400/3600 in a cluster.
How to replace a faulty CRM card on an SRX3400/3600 in a cluster.

If you are using a certificate-based VPN, refer to KB26879 - How to copy certificates after RE replacement in HA cluster [juniper.net] .

Solution

Precautions:

To reduce the risk of affecting the peak usage of traffic, it is recommended to perform any hardware changes during a scheduled maintenance window.

Make sure to follow ESD standards and use the proper tools and procedures, as per the SRX3600 Services Gateway Hardware Guide .

Verification commands:

`show chassis fpc pic-status`	Should show all PICs online for all nodes. (It will take time, especially for SCPs, for PICs to come online after a reboot.)
`show chassis cluster status`	Should show both nodes with the expected priorities for all redundancy groups.
`show system alarms`	Should not show any unexpected alarms.
`show chassis fabric plane`	(High-end only) should show all links "OK".
`request chassis cluster failover redundancy-group {#} node {0\|1}`	Could initiate a failover for cluster configurations to ensure traffic can pass on the system with the new hardware.

Replacing the SPC

The replacement of an SPC card requires that the chassis of the node in which the SPC resides be powered down. You can issue the request system halt operational command and then unplug the power cable. When the system has been powered off, replace the SPC and then power up (or boot the node). To ensure that the new SPC is online, issue the show chassis fpc pic-status command:

root@TEST> show chassis fpc pic-status 
Slot 0   Online       SRX3k SFB 12GE
  PIC 0  Online       8x 1GE-TX 4x 1GE-SFP
Slot 1   Online       SRX3k 16xGE TX
  PIC 0  Online       16x 1GE-TX
Slot 5   Online       SRX3k SPC
  PIC 0  Online       SPU Cp-Flow
Slot 6   Online       SRX3k NPC
  PIC 0  Online       NPC PIC
Slot 7   Online       SRX3k SPC
  PIC 0  Online       SPU Flow

Replacing the NPC

The replacment of an NPC card requires that the chassis of the node in which the NPC resides be powered down. You can issue the request system halt operational command and then unplug the power cable. When the system has been powered off, replace the NPC and then power up (or boot the node). To ensure that the new NPCis online, issue the show chassis fpc pic-status command:

root@TEST> show chassis fpc pic-status 
Slot 0   Online       SRX3k SFB 12GE
  PIC 0  Online       8x 1GE-TX 4x 1GE-SFP
Slot 1   Online       SRX3k 16xGE TX
  PIC 0  Online       16x 1GE-TX
Slot 5   Online       SRX3k SPC
  PIC 0  Online       SPU Cp-Flow
Slot 6   Online       SRX3k NPC
  PIC 0  Online       NPC PIC
Slot 7   Online       SRX3k SPC
  PIC 0  Online       SPU Flow

Replacing the CRM

The replacement of the CRM card requires that the chassis of the node in which the CRM resides be powered down. You can issue the request system halt operational command and then unplug the power cable. When the system has been powered off, replace the CRM and then power up (or boot the node). To ensure that the new CRM is online, issue the show chassis fpc pic-status command:

root@TEST> show chassis fpc pic-status 
Slot 0   Online       SRX3k SFB 12GE
  PIC 0  Online       8x 1GE-TX 4x 1GE-SFP
Slot 1   Online       SRX3k 16xGE TX
  PIC 0  Online       16x 1GE-TX
Slot 5   Online       SRX3k SPC
  PIC 0  Online       SPU Cp-Flow
Slot 6   Online       SRX3k NPC
  PIC 0  Online       NPC PIC
Slot 7   Online       SRX3k SPC
  PIC 0  Online       SPU Flow

Replacing the SFB/RE (Switch Fabric Board) :

The replacement of the SFB/RE card requires that the chassis of the node in which the SFB/RE resides be powered down. You can issue the request system halt operational command and then unplug the power cable. When the system has been powered off, replace the SFB/RE and then power up (or boot the node).

root@TEST> show chassis fpc pic-status 
Slot 0   Online       SRX3k SFB 12GE
  PIC 0  Online       8x 1GE-TX 4x 1GE-SFP
Slot 1   Online       SRX3k 16xGE TX
  PIC 0  Online       16x 1GE-TX
Slot 5   Online       SRX3k SPC
  PIC 0  Online       SPU Cp-Flow
Slot 6   Online       SRX3k NPC
  PIC 0  Online       NPC PIC
Slot 7   Online       SRX3k SPC
  PIC 0  Online       SPU Flow

When the RE is online, confirm if the Junos OS version on it is the same as the earlier version (or same as other node in the cluster):

user@TEST> show version

Model: srx3600
JUNOS Software Release [10.4R4.5] < The version should be the same as the other node in the cluster

If the Junos OS version is not the same, upgrade or downgrade to the required version. For more information, refer to KB16652 - SRX Getting Started - Junos Software Installation/Upgrade [juniper.net] .

The cluster-id is stored in the RE card. When this card is replaced, you will have to reestablish the cluster-id. After the node is powered up, first verify if the cards are online by issuing the show chassis fpc pic-status command. When you have verified that all the cards are online, issue the following command:

set chassis cluster cluster-id X node Y reboot

where X refers to the cluster-id (to which the old RE was configured, and it should match the cluster-id configured on the other node), and Y refers to the node ID of the SFB/RE card being replaced.