Removing and installing a PCIe cable with the power on

Use the following precautions whenever you handle electronic components or cables:

• Attach a wrist strap to an unpainted metal surface of your hardware to prevent electrostatic discharge (ESD) from damaging your hardware.
• If you do not have a wrist strap, before removing the product from ESD packaging and installing or replacing hardware, touch an unpainted metal surface of the system for a minimum of 5 seconds.
• Keep all electronic components in the shipping container or envelope until you are ready to install them.
• If you remove and reinstall an electronic component, temporarily place the component on an ESD pad or blanket, if available.

1. Record the current date and time for use later in the procedure when checking for serviceable events.
2. Determine the identifier string and link identifier for the I/O slot associated with the PCIe cable being replaced by completing the following steps:
   1. Choose one of the following options to view PCIe hardware topology information:
      • If your system is managed by an Hardware Management Console (HMC), complete the following steps:
        1. From the navigation bar, expand Systems Management.
        2. Click Servers and select the server you are working with.
        3. In the Tasks area, expand Hardware Information.
        4. Click PCIe Hardware Topology.
      • If your system is not managed by an HMC, complete the following steps:
        1. Access the Advanced System Management Interface (ASMI) by using an authority level of administrator or authorized service provider. For details about using the ASMI, see Managing the Advanced System Management Interface.
        2. In the ASMI navigation area, expand System Configuration.
        3. Click PCIe Hardware Topology.
   2. Choose one of the following options to determine the link identifier and the I/O slot identifier strings:
      • If one of the links on the PCIe Hardware Topology display has a Link Status of Failed and this is the PCIe cable you want to repair, complete the following steps:
        1. Record the link identifier and the I/O slot identifier string found in the entry that has a Link Status of Failed. The link identifier is in the Link ID column, and the I/O slot identifier string in the I/O Slots column.
        2. Continue with step 3.
      • Find the PCIe cable location code from a serviceable event or error log, by completing the following steps:
        Note: The PCIe cable location code will look similar to Utttt.mmm.sssssss-Px-Cy-Tz. Where tttt is the feature code, mmm is the model, sssssss is the sequence or serial number of the system unit connected by the PCIe cable, x is a number representing the planar position in the system enclosure to which the cable is connected, y is a number representing the card position in the planar, and z is a number representing the PCIe connector position on the card. An example of an actual PCIe cable location code is U2C4B.001.DBJ7805-P1-C2-T1.
        1. Scan the PCIe Hardware Topology data for the entry with a Host Port column value that matches the PCIe cable location code. To identify the location code, see Connector locations.
        2. Record the link identifier and the I/O slot identifier string found in the entry located in the previous step. The link identifier is in the Link ID column, and the I/O slot identifier string in the I/O Slots column.
        3. Continue with step 3.
      • Find the PCIe cable location code from the location code of an I/O slot that connects a PCIe storage enclosure from a serviceable event or error log, by completing the following steps
        Note: The I/O slot location code will look similar to Utttt.mmm.sssssss-Px-Cy-Tz-L1. Where tttt is the type, mmm is the model, sssssss is the sequence or serial number of the system unit connected by the PCIe cable, x is a number representing the planar position in the system enclosure to which the cable is connected, y is a number representing the card position in the planar, and z is a number representing the PCIe connector position on the card. An example of an actual PCIe cable location code is U2C4B.001.DBJ7805-P1-C2-T1-L1.
        1. Scan the PCIe Hardware Topology data for the entry with an I/O Slot column value that matches the I/O slot location code.
        2. Record the link identifier and the I/O slot identifier string found in the entry located in the previous step. The link identifier is in the Link ID column, and the I/O slot identifier string in the I/O Slots column.
        3. Continue with step 3.
      • If you know which PCIe cable you want to repair but do not have the PCIe connector or I/O slot location code, complete the following steps:
        1. Determine the location code of the PCIe connector on the system to which the cable is connected. To identify the location code, see Connector locations.
        2. Scan the PCIe Hardware Topology data for the entry with a Host Port column value that matches the location code of the PCIe connector determined in the previous step.
        3. Record the link identifier and the I/O slot identifier string found in the entry located in the previous step. The link identifier is in the Link ID column, and the I/O slot identifier string in the I/O Slots column.
        4. Continue with step 3.
3. Power off the I/O slot identified in the step 2 by completing one of the follow steps:
   • If the I/O slot is assigned to a logical partition that is powered on and running the Virtual I/O Server (VIOS) or AIX® operating system, complete the following steps:
     1. Log in to the operating system with admin or service level authority. If you need assistance, contact the system administrator.
     2. Choose one of the following options:
        • If the partition is running the VIOS operating system, type diagmenu at the VIOS command prompt and press Enter.
        • If the partition is running the AIX operating system, type diag at the AIX command prompt and press Enter.
     3. On the DIAGNOSTIC OPERATING INSTRUCTIONS display press Enter to continue.
     4. On the FUNCTION SELECTION display, use the arrow keys to select Task Selection (Diagnostics, Advanced Diagnostics, Service Aids, etc.) and press Enter.
     5. On the TASKS SELECTION LIST display, use the arrow keys to select Hot Plug Task and press Enter.
     6. On the Hot Plug Task display, use the arrow keys to select PCI Hot Plug Manager and press Enter.
     7. On the PCI Hot Plug Manager display, use the arrow keys to select List PCI Hot Plug Slots and press Enter.
     8. On the COMMAND STATUS display, locate the entry with the identifier string of the I/O slot in the Slot column, and record the logical device name in the Devices column.
     9. Press F3 (Exit).
        Note: If your terminal emulation does not support F3 to exit, press ESC then type 3 to exit.
     10. On the PCI Hot Plug Manager display, use the arrow keys to select Unconfigure a Device, and press Enter.
     11. On the Unconfigure a Device display, complete the following steps:
         1. Enter the logical device name recorded in step 3.h for the Device Name field.
         2. Use the Tab key to select yes for the Unconfigure any Child Devices field.
         3. Use the Tab key to select no for the KEEP definition in database field.
         4. Press Enter.
     12. On the ARE YOU SURE dialog, press Enter.
     13. On the COMMAND STATUS display, wait for processing to complete and then press F3 twice to return to the PCI Hot Plug Manager display.
     14. On the PCI Hot Plug Manager display, use the arrow keys to select Replace/Remove a PCI Hot Plug Adapter, and press Enter.
     15. On the Replace/Remove a PCI Hot Plug Adapter dialog, use the arrow keys to select the entry with the identifier string of the I/O slot in the Slot column, and press Enter.
     16. On the Replace/Remove a PCI Hot Plug Adapter display, use the Tab key to select the Remove operation, and press Enter.
     17. Follow the online instructions to complete the remove operation. However, do not complete the steps to physically remove the adapter.
   • If the I/O slot is assigned to a logical partition that is powered on and running the IBM® i operating system, complete the following steps:
     1. Log in to the operating system with admin or service level authority. If you need assistance, contact the system administrator.
     2. At the IBM i command prompt, type strsst and press Enter.
     3. Type the service tools user ID and service tools password on the system service tools (SST) Sign On display, and press Enter.
     4. Select the Start a service tool option, and press Enter.
     5. Select the Hardware service manager option, and press Enter.
     6. Select the Packaging hardware resources (systems, frames, cards) option, and press Enter.
     7. On the Packaging Hardware Resources display, locate the entry that contains the I/O slot. To make this identification, scan the display for an entry that has a Description value of System Unit and a Unit ID that matches the U label portion of the I/O slot identifier string. The U label portion is the substring that begins with U and continues until the first dash (-) character.
     8. Type 9 in the Opt column for the system unit entry that you located in step g, and press Enter.
        Note: Specifying option 9 starts the hardware contained within the package view for the selected packaging hardware resource.
     9. On the Packaging Hardware Resources display, locate the entry for the I/O slot. To make this identification, scan the display for an entry with a Location column value that, when appended to the Unit ID value in the upper-right corner, forms the complete I/O slot identifier string.
     10. Type 3 in the Opt column for the entry that you located in step i, and press Enter.
         Note: Specifying option 3 opens the Hardware Resource Concurrent Maintenance menu for the selected packaging hardware resource.
     11. On the Hardware Resource Concurrent Maintenance display, press F9 to start the power-off domain function.
     12. Follow the online instructions to complete the power-off operation.
   • If the I/O slot is assigned to a logical partition that is powered on and running the Linux operating system, complete the following steps:
     1. Log in to the operating system with admin or service level authority. If you need assistance, contact the system administrator.
     2. At the Linux command prompt, type drmgr -c pci -r -s io_slot_identifier, where io_slot_identifier is the identifier string of the I/O slot, and press Enter.
     3. Follow the online instructions to complete the remove operation. However, do not complete the steps to physically remove the adapter.
   • If the I/O slot is not assigned to a logical partition that is powered on, continue to the next step.
4. Activate the identify indicator for the PCIe connector to which the PCIe cable is connected by completing the following steps:
   1. Choose one of the following options:
      • If your system is managed by an HMC, complete the following steps:
        1. From the navigation bar, expand Systems Management.
        2. Click Servers and select the server you are working with.
        3. In the Tasks area, expand Hardware Information.
        4. Click PCIe Hardware Topology.
      • If your system is not managed by an HMC, complete the following steps:
        1. Access the ASMI by using an authority level of administrator or authorized service provider. For details about using the ASMI, see Managing the Advanced System Management Interface.
        2. In the ASMI navigation area, expand System Configuration.
        3. Click PCIe Hardware Topology.
   2. Scan the PCIe Hardware Topology data for the entry with the link identifier and I/O slot identifier recorded in step 2.
   3. Select the PCIe link and click Identify Indicators.
   4. On the Identify Indicators display, select both locations and click Activate LED.
5. If the system has a rear cover, remove or open it.
6. Remove and replace the PCIe cable by completing the following steps:
   1. Locate the PCIe connector on the system with the active identify indicator and disconnect the PCIe cable from this located PCIe connector.
   2. Disconnect the other end of the PCIe cable from the PCIe storage enclosure.
   3. Connect the replacement PCIe cable to the PCIe connector from which the faulty PCIe cable was removed in step 6.a.
   4. Connect the other end of the replacement PCIe cable to the PCIe storage enclosure from which the faulty PCIe cable was removed in step 6.b.
7. If the system has a rear cover, close it.
8. Deactivate the identify indicator that you activated in step 4 by repeating step 4 as specified, except to click Deactivate LED.
9. Power on the I/O slot identified in the step 2 by choosing one of the following options:
   • If the I/O slot is assigned to a logical partition that is powered on
   • If the I/O slot is not assigned to a logical partition that is powered on
   If the I/O slot is assigned to a logical partition that is powered on

   Note: If your system is not managed by an HMC and is powered on to firmware running state, choose this option.

   • If the I/O slot is assigned to a logical partition that is powered on and running the Virtual I/O Server (VIOS) or AIX operating system, complete the following steps:
     1. Log in to the operating system with admin or service level authority. If you need assistance, contact the system administrator.
     2. Choose one of the following options:
        • If the logical partition is running the VIOS operating system type diagmenu at the VIOS command prompt, and press Enter.
        • If the logical partition is running the AIX operating system type diag at the AIX command prompt, and press Enter.
     3. On the DIAGNOSTIC OPERATING INSTRUCTIONS display, press Enter to continue.
     4. On the FUNCTION SELECTION display, use the arrow keys to select Task Selection (Diagnostics, Advanced Diagnostics, Service Aids, etc.), and press Enter.
     5. On the TASKS SELECTION LIST display, use the arrow keys to select Hot Plug Task, and press Enter.
     6. On the Hot Plug Task display, use the arrow keys to select PCI Hot Plug Manager, and press Enter.
     7. On the PCI Hot Plug Manager display, use the arrow keys to select Add a PCI Hot Plug Adapter, and press Enter.
     8. On the Add a PCI Hot Plug Adapter dialogue, use the arrow keys to select the entry with the identifier string of the I/O slot in the Slot column, and press Enter.
     9. Follow the instructions that display to complete the add operation. However, do not complete the steps to physically add the adapter.
     10. Press F10 twice to return to the AIX command prompt.

     11. For an AIX logical partition: At the command prompt, type, cfgmgr -l logical_device_name, where logical_device_name is the logical resource name of the I/O slot recorded in step 3.h.

         For a VIOS logical partition: At the command prompt type cfgdev -dev logical_device_name, where logical_device_name is the logical resource name of the I/O slot recorded in step 3.h.

   • If the I/O slot is assigned to a logical partition that is powered on and running the IBM i operating system, complete the following steps:
     1. Log in to the operating system with admin or service level authority. If you need assistance, contact the system administrator.
     2. At the IBM i command prompt, type strsst and press Enter.
     3. Type the service tools user ID and service tools password on the system service tools (SST) Sign On display, and press Enter.
     4. Select the Start a service tool option, and press Enter.
     5. Select the Hardware service manager option, and press Enter.
     6. Select the Packaging hardware resources (systems, frames, cards) option, and press Enter.
     7. On the Packaging Hardware Resources display, locate the entry that contains the I/O slot. To make this identification, scan the display for an entry that has a Description value of System Unit and a Unit ID that matches the U label portion of the I/O slot identifier string. The U label portion is the substring that begins with U and continues until the first dash (-) character.
     8. Type 9 in the Opt column for the system unit entry that you located in step 9.g, and press Enter.
        Note: Specifying option 9 starts the hardware contained within the package view for the selected packaging hardware resource.
     9. On the Packaging Hardware Resources display, locate the entry for the I/O slot. To make this identification, scan the display for an entry with a Location column value that, when appended to the Unit ID value in the upper-right corner, forms the complete I/O slot identifier string.
     10. Type 3 in the Opt column for the entry that you located in step 9.i, and press Enter.
         Note: Specifying option 3 starts the Hardware Resource Concurrent Maintenance menu for the selected packaging hardware resource.
     11. On the Hardware Resource Concurrent Maintenance display, press F10 to start the power-on domain function.
     12. Follow the online instructions to complete the power-on operation.
   • If the I/O slot is assigned to a logical partition that is powered on and running the Linux operating system, complete the following steps:
     1. Log in to the operating system with admin or service level authority. If you need assistance, contact the system administrator.
     2. At the Linux command prompt, type drmgr -c pci -a -s io_slot_identifier, where io_slot_identifier is the identifier string of the I/O slot, and press Enter.
     3. Follow the online instructions to complete the add operation. However, do not complete the steps to physically add the adapter.
   If the I/O slot is not assigned to a logical partition that is powered on

   Note: If your system is not managed by an HMC and is powered on to firmware standby state, choose this option.
   Choose one of the following options:

   • If your system is managed by an HMC, complete the following steps:
     1. From the navigation bar, expand Systems Management.
     2. Click Servers and select the server you are working with.
     3. In the Tasks area, expand Hardware Information.
     4. Click PCIe Hardware Topology.
   • If your system is not managed by an HMC, complete the following steps:
     1. Access the ASMI by using an authority level of administrator or authorized service provider. For details about using the ASMI, see Managing the Advanced System Management Interface.
     2. In the ASMI navigation area, expand System Configuration.
     3. Click PCIe Hardware Topology.
     4. Scan the PCIe Hardware Topology data for a value in the I/O Slots column that matches the identifier string of the I/O slot, and select the entry.
     5. Click Probe Link.
        Important: If the Probe Link function fails, verify that the I/O slot is not owned by a partition that is powered on. Then choose from the following options:

        • If you determine that the I/O slot is owned by a partition that is powered on, return to step 9 and choose the alternative option that supports an I/O slot assigned to a partition that is powered on.
        • If you are certain the I/O slot is not owned by a partition that is powered on, contact your next level of support for assistance.
10. Verify the PCIe hardware topology by completing the steps in Verifying the PCIe hardware topology.

    All PCIe hardware topology problems that are discovered in this step must be corrected before continuing with this procedure.

11. Check for and handle new serviceable events that were generated during the procedure by completing the steps in Checking for new serviceable events. Use the date and time information that you recorded in step 1 as the start time for searching for new serviceable events.