Installation and Migration Guide

Replacing a node with an equivalent node

Follow the steps in this section to replace a node in one of your frames with an equivalent node. Equivalent nodes have the same adapters and configuration data, and have the same node type. Note that if you are replacing a node with another node, you must also ensure that the microcode level of the node is correct. Refer to Step 34: Update the state of the supervisor microcode for more information.

Replacing a node with an equivalent node means removing one node and adding another in its place. Before you do this, you need to determine what applications are currently running on the node. You may need to plan to transfer these applications to another node on your system. Also, if the node you want to delete serves as the boot/install server or as a primary node, you may need to transfer the responsibilities to one of the remaining nodes.

• Be sure to back up any local data you want to preserve.
• Do these steps in the system partition in which you replace the nodes (SP Switch only).

Step 1: Shut down the node and power it off

Prior to shutting down the node, make a list of all the applications and programs running on the node (such as LoadLeveler) and stop them from running.

Use Perspectives or the cshutdown command to shut down the nodes. For example, issue:

cshutdown -G -N 48

|Step 2: Security considerations

|Step 2.1: Secure remote commands

|If running with a secure remote command method enabled, changing the host |name or IP address might require a regeneration of public keys and |known_hosts files, depending on your secure remote command |configuration.

|Step 2.2: Unconfigure DCE-related information for the node (required for DCE)

|After the node has been deleted from the frame, you must remove any |DCE-related principles and objects from the DCE registry. You must also |unconfigure DCE (Admin portion). You must have cell administrator |authority to perform this step.

|On the control workstation, use the rm_spsec -t admin |node_dce_hostname command first, then do a DCE Admin unconfigure for |the node (smit rmdce).

|To remove any additional principals related to the node using the SMIT |panels, enter the host name of the principals to be deleted. For |example, on the "Admin only unconfiguration for another machine" panel |in the "Machine's name or TCP/IP address" field, enter the host |name for the additional adapters.

|For the nodes being removed, verify that all DCE principals have been |deleted from the DCE registry. Issue:

|dcecp -c principal catalog -simplename

|You must now issue:

|setupdce -v
| 
|config_spsec -v

|Note:

To run rm_spsec, setupdce, and config_spsec off of |the SP, you must set the SP_NAME environment variable on the remote |workstation to point to the SDR of the SP system being reconfigured. In |addition, rm_spsec and config_spsec may require the |-r (remote) flag. Refer to the rm_spsec and |config_spsec commands in PSSP: Command and Technical |Reference for a description of the -r (remote) flag. |

Step 3: Replace the old node with the new one and power it on

Your IBM Customer Engineer (CE) performs this step.

When the node is powered on, the node supervisor contacts the frame supervisor and informs it that it is running. The frame supervisor updates the hardmon daemon, running on the control workstation, that a new node has been added to the system.

Step 4: Update the state of the supervisor microcode

To ensure that you have the latest level of microcode required by the hardware on your SP system, issue the spsvrmgr command. For example, to get the status in report form of all of your frames, nodes, and switches, enter:

spsvrmgr -G -r status all

To update the microcode of the frame supervisor of frame 3, enter:

spsvrmgr -G -u 3:0

Note:

When using the spled application, a node in the process of having the microcode on its supervisor card updated will not be displayed in the window.

Refer to the PSSP: Command and Technical Reference for more information on using the spsvrmgr command.

Step 5: Unallocate NIM resources

Unallocate NIM Resources on the boot/install server for the node you are replacing by issuing the unallnimres command. For example:

unallnimres -l 33

Step 6: Delete the NIM client

Delete the NIM client on the boot/install server for the node you are replacing by issuing the delnimclient command. For example, to delete the NIM client for node number 3, issue:

delnimclient -l 3

Step 7: Acquire the hardware Ethernet address

This step takes the hardware Ethernet address for your new node (either from the node itself or from a file), puts it in the Node Object of the SDR, and sets up the /etc/bootptab file on the boot/install server.

Note:

If you have information for the node being replaced in /etc/bootptab.info, be sure to remove the old address and then add the new Ethernet address.

Use the sphrdwrad command to write the hardware Ethernet address to the SDR. For example:

sphrdwrad -l 3

Step 8: Set up nodes to be installed

Use SMIT or issue the spbootins command to change the default boot/install information in the Node Objects in the SDR so that you can indicate a different bootp response for nodes to be installed.

For example:

spbootins -s yes -r install -l 3

This step also runs the setup_server command on all affected server nodes so that the nodes served are installed accordingly at power-up.

Note:

|In order to run the spbootins -s yes command and be |authorized to perform a remote command to the target nodes, you must have SDR |write authority and be authorized to perform an rsh to the target |nodes. Therefore, your user ID must be in the appropriate authorization |file (.k5login, .klogin, or |.rhosts) on the target nodes. |

Step 9: Network boot the new SP node

You can network boot the nodes using the following command:

nodecond frame_id slot_id &

Note:

For MCA nodes, the nodecond command remotely processes information from the initial AIX firmware menus. You should not change the language option on these menus. The language must be set to English in order for the nodecond command to run properly.

Step 10: Run post-installation procedures

Set up the new nodes for any additional tools you use to manage your RS/6000 SP system environment.

See Table 4 for a listing of information on the procedures for setting up these facilities.