H13697 - RAID ARRAY FIXED DISK DRIVE REPLACEMENT PROCESS TEXT: The following procedure was developed to assist in implementing fixed disk replacement under 8641 ECA002, which provides for the replacement of specific 2.25 GB fixed disk drives. (Refer to RETAIN record # H13646 for important additional information.) However, this process may prove to be beneficial at any time multiple fixed disks are being replaced in a RAID Array. VERY IMPORTANT NOTES: NOTE 1: The most important points of this procedure are: a. The Rebuild Procedure is only for RAID 1 and RAID 5 configurations. For RAID 0 configurations, all data must be backed up to tape. Then, after replacing the drives, a a restore from tape will be required. b. Keeping track of which drives are located in which bays and, c. To have a backup copy of the customer's RAID configuration. NOTE 2: The drives may have been swapped out quite a bit. In the event of drive failure, inspect the connectors for damage. NOTE 3: If the fixed disks are being replaced due to suspected STICTION problems, the MAIN IDEA IS NOT TO POWER DOWN THE SERVER. If the IBM F/W Streaming RAID Adapter Option Diskette is at version 2.3 it is STRONGLY RECOMMENDED that version 2.23. be obtained and used during this procedure. Version 2.3 TURNS OFF ALL THE FIXED DISKS in the Array during the rebuild process. This will expose the remaining drives to possible stiction. Reinstall version 2.3 after all drives have been replaced. "REBOOT," in this procedure refers to soft boots (CTRL-ALT-DEL). 1. Make a Chart of Bays 1 through 6 and write in each bay the original disk drive Serial Number and the New disk drive serial number for that particular bay. 2. While still installed in the system, place stickers on the drives indicating what bays they are in. Do the same to the new drives once they are installed. 3. Soft Boot the IBM F/W Streaming RAID Adapter Option Diskette Version 2.23. (DO NOT USE V2.3. See important note above). Select "Initialize/Synchronize Array." Select Synchronize Logical Drive. Press the spacebar to select each logical drive. Once all logical drives are highlighted, Press enter to begin synchronization. 4. After the synchronization completes, Return to the Main Menu. Select "Advanced Options." Then, select "Backup Configuration to Diskette" from this menu. Insert a blank diskette into drive A: and put in a filename for this backup. The default filename is "config." If you use the same backup diskette for more than one system, be sure to give each system a different filename and keep track of the specific filenames. 5. Exit out of Advanced Options. Select "View Configuration" from the "Main Menu." Identify which bay contains the Hot Spare drive, if applicable. 6. Pull the Hot Spare drive out of its bay. Exit out of "View Configuration" and the "Main Menu." 7. Reboot, Press Y to update the configuration with the HSP bay as "DDD". 8. You are now ready to pull out the first drive to be replaced The first drive to replace should be the drive SUSPECTED of having the highest probability of stiction. If there is not a SUSPECT drive, then choose an order you would like to replace them in, such as Bays 1, 2, 3, 4,etc. Just remember what order that is. Checking off the bays on your chart after they have completed is very helpful in keeping track. NOTE: Remember, the HOT SPARE drive is pulled out first and is left out during replacement/rebuilding of the other drives. It then becomes the last drive replaced and its status turns to HSP. The HOT Spare doesn't require rebuilding (it's a spare) In order you have chosen, follow steps 8 and 9 for each drive. 9. Pull out Drive and reboot the RAID Support Diskette. "Y" to update configuration as DDD. Reboot. Install new drive into the bay. Select "Rebuild Drive" off the "Main Menu." Then, select "Replace Drive." Highlight correct DDD Drive in the bay you are replacing. The system will start the drive. After starting successfully, the drive status will change from DDD to OFL. 10. Select "Rebuild Drive" on the Rebuild Menu off the Main Menu Highlight correct OFL Drive in the bay you are rebuilding. System will rebuild the drive from 0 to 100%. After it has completed successfully, it will change the drive from OFL to ONL. IF SOFT ERRORS OCCUR, DO THE FOLLOWING: 1. Before rebooting, select Drive information to determine which drive had Soft Errors. Take note of the drive that had Soft errors. Note: Soft errors initialize to zero upon reboot. 2. The Quantum drive that failed to rebuild will display OFL status Replace this Quantum with the starfire drive that was originally in this bay. 3. Restore the configuration that was backed up in Step 4 by selecting Advanced Functions on Main Menu. Then, choose Restore Configuration from Diskette. Enter the filename of the saved configuration for this server. 4. Once the RAID configuration has completed being restored, all drives should indicate ONL. At this time, remove the drive that was noted as causing Soft Errors. 5. Reboot to RAID utility. Say "Yes" to update the RAID configuration with the Empty Bay as DDD. When the update completes, the RAID utilty will prompt you to Reboot the system to the RAID utility. 6. Install the Quantum in the DDD bay and rebuild it. When this drive completes rebuilding, your Soft Error drive will be replaced. You are now past the Soft Errors and can begin replacing the rest following the procedures. IF FATAL ERROR OCCURS DO THE FOLLOWING: Replace only the original drive back into the system that has not been rebuilt successfully yet. Then, restore the configuration from diskette by rebooting to Option Diskette. Select "Restore Configuration" from diskette on the Advanced Options Menu off the Main Menu. Put the diskette with the backup configuration into the diskette drive and type in the filename for that particular system. Then, proceed replacing and rebuilding drives wherever you left off. If during the restoring from diskette an, "ADAPTER IS NOT RESPONDING TO CONTROLLER COMMANDS" message appears, do the following: Reboot and Restore the configuration from the backup RAID Support Diskette again until it comes back and says restoration completed successfully. 11. Once all drives are replaced and rebuilt successfully, power off the server and Cold Boot the system. The system should now be ready to return to the customer.