Connecting and configuring the disk drive enclosure in a system running Linux

Use the information in this section to perform the cabling, SCSI addressing, and postconnection requirements for the disk drive enclosure.

1. Install the initiator repeater cards into the disk drive enclosure. For instructions, see SCSI repeater card.
2. Connect the SCSI cable to your repeater card. Choose one of the following options based on your situation:
   • If you are using only single repeater cards or only dual repeater cards, connect your SCSI cable (A) to the initiator repeater card (B). For more details, see Figure 1, Figure 2, or Figure 3, depending on your situation.
     Figure 1. SCSI cable to a single repeater card

     
     Figure 2. SCSI cable to a dual repeater card that is connected with two SCSI cards

     
     Figure 3. SCSI cable to a dual repeater card that is connected with four SCSI cards

     
   • If you are using a combination of single and dual repeater cards, connect a SCSI cable to the dual-initiator repeater card (A). Then, connect the dual repeater card (A) to the single repeater card (B) by using a different SCSI cable. For more information, see Figure 4.
   Figure 4. SCSI cable to a dual repeater card that is connected to a single repeater card

   
3. Complete the SCSI cable connection process in this step.
   Important: The SCSI adapter or the system or partition must be powered off before making the connection in this step.

   Connect the other end of the SCSI cable (C) to the SCSI adapter (D) on the server. For more details, see Figure 5.

   Figure 5. SCSI cable to a SCSI interface card

   
4. Add the disk drives to the operating system that you are using. See Disk drives and solid state drives.
   Important: If you plan on using the disk enclosure in a Linux environment, it is critical that you set up a Linux CRON job so that if errors exist, they are reported to the attached system. The Linux CRON job must be set up during the initial installation.
5. Set up the Linux CRON job during the initial installation of your SCSI disk drive enclosure. Perform the following tasks to ensure that your Linux CRON job is set up:
   1. Install the following RPMs from Service and productivity tools (http://www14.software.ibm.com/webapp/set2/sas/f/lopdiags/home.html):
      • librtas
      • lsvpd
      • ppc64-utils
      • diagela
   2. If your logical partitions are managed by an HMC, install the src, rsct.core.utils, rsct.core, csm.core, csm.client, and devices.chrp.base.ServiceRM packages in order to report errors to the HMC. Instructions for installing the packages are located at Service and productivity tools (http://www14.software.ibm.com/webapp/set2/sas/f/lopdiags/home.html).
      Note: Errors on hard disks within the enclosure are not found using this CRON job. This diagnostic test is limited to reporting errors with the fans, power supplies, Vital Product Data (VPD) card, and repeater in the enclosure.
   3. To run enclosure diagnostics, add the diag_encl job to the system CRON table. As root user, edit the system CRON by using the crontab -e command. At the bottom of that file, enter:

      * 3 * * * /usr/sbin/diag_encl -s 1>/dev/null 2>&1

      The contents of the script depend on the configuration of the system to which the enclosure is attached.

      This CRON runs the diagnostic application diag_encl daily at 3 a.m. The application reads the contents of the /etc/diagela/diagela.config file to determine which applications have been registered to be notified of errors. To add additional notification methods or for more information, see that file. By default, the following notifications occur:

      • If the system is managed by an HMC, notifications are shown in the Manage Serviceable Events task on the HMC.If the system is not managed by an HMC, the root group is notified.
      • If the system has the Electronic Service Agent™ application installed, it is notified.
      • The error details are printed to the console, to the end of the /var/log/platform log file, and to the syslog file in /var/log/messages.
        Notes:

        1. On a logically partitioned system, it is only necessary to have the CRON job on one partition that is associated with the enclosure. However, it is acceptable to have the same CRON job on all partitions.
        2. The diag_encl application might run at any time to diagnose enclosures that are attached to the system.