How to remove and replace a Starcat PCI Card:ATR:1527:2

Asset ID:	1-71-1310146.1
Update Date:	2016-01-20
Keywords:

Solution Type Technical Instruction Sure

Solution 1310146.1 : How to remove and replace a Starcat PCI Card:ATR:1527:2

Applies to:

Sun Fire E20K Server - Version: Not Applicable and later   [Release: N/A and later ]
Sun Fire E25K Server - Version: Not Applicable and later    [Release: N/A and later]
Sun Fire 12K Server - Version: Not Applicable and later    [Release: N/A and later]
Sun Fire 15K Server - Version: Not Applicable and later    [Release: N/A and later]
Information in this document applies to any platform.

Goal

How to remove and replace a Starcat PCI Card
********************************************************************************
To report errors or request improvements on this procedure,
please go to http://support.us.oracle.com and put a comment on Doc ID: 1310146.1
********************************************************************************

Solution

CAP PROBLEM OVERVIEW: PCI Failure

DISPATCH INSTRUCTIONS

WHAT SKILLS DOES THE ENGINEER NEED:
Starcat Training

Task Complexity: 2

FIELD ENGINEER INSTRUCTIONS

WHAT STATE SHOULD THE SYSTEM BE IN TO BE READY TO PERFORM THE RESOLUTION ACTIVITY? :

A data backup is not a pre-requisite but is a wise precaution.
The customer is not required to perform a system shutdown.
Someone with root access is required to properly unconfigure/configure the card per instruction below.
Customer should study the possibility of replacement by DR:

Verify for each I/O adapter being detached that an alternate path to system critical resources on another
I/O board exists.

WHAT ACTION DOES THE ENGINEER NEED TO TAKE:

DR:

DR Detach:

NOTE: Pulling card without properly disconnecting from Solaris may cause domain to panic if doing an active replacement.

Ensure that the adapter in the cassette to be removed has terminated all input/output operations
# cfgadm -v -c disconnect AP-ID

Wait for the amber removal indicator LED to illuminate
Label and remove the I/O cables from the cassette.

Remove PCI Cassette
Replace PCI Card.

NOTE: If replacing a Fibre Channel (FC) HBA, make note of WorldWide Numbers (WWN) for each port on new card and provide wwn/port to customer.

Insert he PCI cassette into the appropriate slot (C5V0, C3V0, C3V1, or C3V2)
Reconnect the I/O cables to the cassette.

DR Attach:
cfgadm -v -c configure AP-ID

No DR:
Shut down Domain
Power off Domain
Replace PCI Card following service manual procedure.

NOTE: If replacing a Fibre Channel (FC) HBA, make note of WorldWide Numbers (WWN) for each port on new card and provide wwn/port to customer.

Power on domain and Hpost lavel 32
Start the domain

If FC HBA card, connect Loopback plug into port(s) and have customer verify "online" status of correlating fc hba card port(s) by running following command:

NOTE: If you do not have an external Loopback plug, you can make one by using a known good LC multimode fiber cable and taking it apart and plugging the same cable (Fibre Strand) into the transmitter and receiver of the fc hba port. Many LC cables have standalone connectors on each strand with the ends held together with a clip making it easy to separate, others have ends molded together, in which case they can be cut apart using a scissor or knife.

# fcinfo hba-port
If ports do not list as "online" then fc hba card maybe DOA. Work with Oracle SAN support to address any issues before leaving the site.

OBTAIN CUSTOMER ACCEPTANCE

WHAT ACTION DOES THE CUSTOMER NEED TO TAKE TO RETURN THE SYSTEM TO AN OPERATIONAL STATE:

Changing a HBA with an embedded WWN may require SAN re-configuration.
Customer "MAY" need to update any FC Switch zoning and/or FC Storage configurations with new WorldWide Port Numbers (WWPN) from each port on new card. This is a customer administrative activity per standard Oracle contract.

Refer customer to following doc for troubleshooting assistance if needed.
SAN: Troubleshooting Solaris[TM] 10 (not for Solaris[TM] 8 & 9) Fibre Channel (FC) HBA list in "NOT CONNECTED" status. (Doc ID 1303745.1)

Check status of multi-path software

PARTS NOTE:

N/A

REFERENCE INFORMATION:
Service Manual: http://download.oracle.com/docs/cd/E19065-01/servers.e25k/index.html

TIME ESTIMATE: 120 minutes

Attachments

This solution has no attachment