Replacement Procedure

4 Control Processor Blade (CP8) Replacement Procedure

53-1001806-01

• “Slot unknown” message relating to a CP slot

• CP blade errors or I

C timeouts

• FRU: FRU_FAULTY messages for a CP blade

• Configuration loader messages or “Sys PCI config” messages

• Generic system driver messages (“FABSYS”)

• Platform system driver messages (“Platform”)

• EM messages that indicate a problem with a CP blade

• Function fail messages for the CP master

For more information about error messages, refer to the Fabric OS Message Reference.

Recording critical backbone information

Back up the backbone configuration before you replace a CP blade. Refer to the Fabric OS Administrator’s Guide for

backup information.

ATTENTION

The instructions below reference specific slot numbers. these numbers will be different between the PowerConnect

B-DCX-4S and PowerConnect B-DCX.

1. Log in to the functioning CP blade as admin, using the serial console connection.

2. Type haShow to determine which CP blade is active:

swDir:admin> haShow

Local CP (Slot 7, CP1) : Active

Remote CP (Slot 6, CP0) : Standby, Healthy

HA Enabled, Heartbeat Up, HA State Synchronized

3. If the functioning CP blade is performing as the active CP blade, go to step 4. If the faulty CP blade is performing

as the active CP blade, fail over the blades:

a. Log in to the faulty CP blade as admin, using a serial console connection.

If you connect successfully to the faulty CP, continue to step b.

If you cannot connect to the faulty CP, remove the faulty CP blade for repair as follows:

• Turn the CP blade off and notify the backbone of a hot swap request by sliding the slider switch in the

top ejector down to the off position (Figure 1).

• Wait until the active CP LED on the functioning CP blade is lit. Remove the faulty CP blade.

b. Type haFailover. The functioning CP blade becomes the active CP blade.

c. Wait until the status LED on the functioning CP blade is no longer lit.

d. Type haShow from the functioning CP blade (the new active CP blade) to verify the failover.

Depending on the failure mode of the faulty CP blade, the haShow output may not show the “HA State

Synchronized” message but may indicate that the system is in non-redundant mode.