Cleaning Up the System Resource Management (SRM) Database

Resolving The Problem

How can I remove (clean up) a failed or unwanted hardware resource that is listed in the SRM database? How can I change the resource name for valid hardware to start from the beginning? This document uses one example for removing Local Work Station (LWS) and changing the resource name of an existing LWS. The process is the same for other resources.

Caution: SRM is system managed. If the command GO RESTORE Option 21 or RSTCFG *ALL (which restores ALL configurations saved using GO SAVE Option 21) is issued from a different model of IBM iSeries 400, the SRM database will be overwritten. As a result, it can contain entries that are not valid; the hardware will report as FAILED. Therefore, you should ensure the RSTCFG command has parameter SRM set to *NONE when performing a full system restore to Model B; for example, from a full system save done on Model A. The default is *ALL, see below for more details.

The following information is from the i5/OS online help about Restore Option 21:

Select this option to restore the system other than the i5/OS licensed program. This option performs the following operations: o Changes the system operator (QSYSOPR) message queue to *BREAK or *NOTIFY delivery mode. o Ends all subsystems. o Restores: - User profiles - Configuration objects - All IBM-supplied libraries including those containing user data - All user libraries - Folders, documents, and mail - All objects in directories - Private authorities o Start controlling subsystem This option runs the following commands: ENDSBS SBS(*ALL) OPTION(*IMMED) RSTUSRPRF RSTCFG OBJ(*ALL) RSTLIB SAVLIB(*NONSYS) RSTDLO DLO(*ALL) SAVFLR(*ANY) RST OBJ(('/*') ('/QSYS.LIB' *OMIT) ('/QDLS' *OMIT)) RSTAUT USRPRF(*ALL) STRSBS SBSD(controlling-subsystem)

As shown here, the default parameters are normally used for the RSTCFG command. In this case, it indicates the parameter value for SRM is *ALL. In addition, if no changes are made, the SRM database is copied from Model A to Model B. However, as stated in the appendix of the Backup and Recovery guide on recovering to a different server, restore to a different system should be changed to 'Y' on the GO RESTORE Option 21 which will set the RSTCFG to have SRM(*NONE).

The methods are listed below to recover a SRM database that is damaged, altered, incorrect, and so on. One method is to perform a scratch installation. Then, restore the user data following the iSeries Backup and Recovery Version 5 (SC41-5304-06). Another method is to perform a manual cleanup of the SRM database.

Method 1:

Note: Follow the IBM OS/400 Backup and Recovery guide for the complete steps. The Save/Restore queue can assist with providing documents under iGSC support. For walk-through assistance, a consulting agreement is required.

STEP	ACTION/COMMENT
1.	Back up user data only on new system.
2.	Scratch install the system with invalid SRM data.
3.	Restore from the last GO SAVE Option 21 media, but with the SRM parameter set to *NONE Restore the configuration objects on the target system using the Restore Configuration (RSTCFG) command. Specify *NONE for the System Resource Management. Once the objects are restored, some changes might be required. For example, the controller and line types might be different. Delete them and create them again with the correct type and resource. To restore the configuration objects, on the IBM OS/400 or IBM i5/OS command line type the following: RSTCFG OBJ(*ALL) DEV(tape-device-name) SRM(*NONE) ALWOBJDIF(*ALL)
4.	Restore user data, and make the necessary changes.

Method 2:

Note: An IBM Support Specialist can perform the steps listed in Method 2 under a consulting agreement only.

STEP	ACTION/COMMENT
1.	Perform manual cleanup, and make changes to resources and configuration objects: a. Change the IPLTYPE system value to 2. b. IPL
2.	STRSST a. Select Option 1, Start a service tool, and press the Enter key. b. Select Option 7, Hardware service manager, and press the Enter key. c. Select Option 2, Logical hardware resources (buses, IOPs, controllers,...), and press the Enter key. d. Select Option 1, System bus resources, and press the Enter key. e. Press F9=Failed resources. f. Delete all resources with failed status. Do not delete any resources if the message No failed logical hardware resources were found is received. g. You may want to print (F6=Print) the Logical Hardware Resources on System Bus for reference; a QPCSMPRT spooled file is created. Back out to the OS/400 or i5/OS command line.
3.	Run the WRKHDWRSC TYPE(*LWS) command, and write down or print the screen. For example: Resource Type Status Text CMB01 284C Operational Combined function IOP CTL01 2746 Operational Work Station Controller CTL03 6A59 Not detected Work Station Controller On the Operational resource, select Option 5, Work with controller descriptions and press the Enter key. CTL01 in this example. Write down or print the Controller Descriptions screen, example: CTL01 CTL99 QCTL
4.	Return to STRSST: a. Select Option 1, Start a service tool, and press the Enter key. b. Select Option 7, Hardware service manager, and press the Enter key. c. Select Option 2, Logical hardware resources (buses, IOPs, controllers,...), and press the Enter key. d. Select Option 1, System bus resources, and press the Enter key. e. On the resource name you want to change, select take option 2. Change detail, and press the Enter key.. For this example, it is CMB01 that is changing to CMB88. f. Type the New resource name (it must be unique), and press the Enter key. Write down or print the screen with the new name (CMB88). g. On the Logical Hardware Resources on System Bus screen, select Option 9, Resources associated with IOP on the same Resource Name, and press the Enter key. h. On the Logical Hardware Resources Associated with IOP screen, select Option 2, Change detail on the Resource Name you want to change, and press the Enter key. For this example, CTL01 is changing to CTL88. i. SST.
5.	Run the WRKCTLD CTLD(CTL88) command. a. Select Option 2, Change on CTL88, and press the Enter key. Change the Resource name value from CTL01 to CTL88.
6.	Repeat the above process (beginning with Step 3) for each hardware resource.
7.	Once all the changes have been made, do the following: a. Change system value QIPLTYPE to 0=Unattended IPL. b. IPL the system in normal mode.