Dear Customer, Enclosed is Support Level Supplement (SLS) unx222. This floppy contains a device driver for the following Irwin tape drives running under SCO UNIX System V/386 Operating System (GT/MC) for the following release: SCO UNIX System V/386 Release 3.2.0 Operating System Open Desktop Release 1.0 SCO UNIX System V/386 Release 3.2 Operating System Version 2.0 Supported Irwin Tape Drives =========================== 110 225 2020 125 245 2040 145 285 2080 787 2120 Supported Controllers ===================== Primary floppy disk controller Irwin 4251 drive multiplexor board Irwin 4100 bus controller card (For use with the 787 and 2120 drives) Irwin Tape Driver Installation Procedure ======================================== Place the Irwin driver diskette (SLS unx222) in your floppy drive and type the following at a root prompt: cd / /etc/custom From the custom menu pick: Install -> A New Product -> Entire Product -> Continue Next, you will see the messages: Installing custom data files. Creating file list. Extracting files. Executing Unix Irwin Tape Driver Init Script. Next, you will see the message: The Irwin Tape Driver is configured. You must relink the kernel and reboot the system in order to make the driver active. Do you want to Re-Link the Kernel now (y/n) The Irwin tape driver has now been extracted. To actually link the driver into your SCO UNIX System V/386 kernel, answer "y" and then enter . Next you will see: The UNIX Operating System will now be rebuilt. This will take a few minutes. Please wait. Root for this system build is /. After a few minutes you will see the following message: The UNIX Kernel has been rebuilt. You will then be asked: Do you want this kernel to boot by default? (y/n) Again, answer "y" and then enter . The next system message you will see is: Backing up /unix to unix.old Installing new /unix The kernel environment includes device node files and /etc/inittab. The new kernel may require changes to /etc/inittab or device nodes. You will then be asked: Do you want the kernel environment rebuilt? (y/n) Again, answer "y" and then enter . The next system message you will see is: Setting up new kernel environment Next you will be taken back to Custom with the prompt: Press any key to continue After pressing any key, you may quit Custom. You should now create an emergency Boot/Root floppy with the appropriate tape device on it. This is vital in the case of a system crash where you need to boot from your emergency Boot/Root floppy and restore data from your backup tape. To do this at the prompt, type: mkdev fd At the menu, pick the appropriate option for the type of floppy you are using. NOTE: You cannot use 48tpi. At the next menu, pick the option for: Bootable only followed by the option for: Root filesystem only This is because 96ds15 and 135ds9 floppies cannot fit the amount of data needed for a single Root/Boot floppy. In either case, you will be asked: Would you like to format the floppy first? (y/n) In case your floppy is unformatted. When this process is complete, remove the floppy from the drive and store it in a safe place. In the Event of a system crash ============================== The emergency Boot/Root floppy that has been prepared for you contains all you will need to boot up and issue a restore command after a system crash. When you boot up using your emergency floppy, you will be put in System Maintenance mode at a root prompt. You can then restore your backup using the following two commands: /etc/mcdaemon /bin/restore rf /dev/rctmini /dev/hd0root The "/etc/mcdaemon" command starts a daemon process which is necessary to the operation of the Irwin tape drive. When you boot up from the hard disk this daemon is automatically started for you, but when you boot up from your Boot/ Root floppy, you need to issue the "/etc/mcdaemon" command before your restore command. The restore command itself restores a root filesystem backup from the Irwin tape device. This command is to be used for tape backups made using the "backup", or "sysadmin" utilities. The "mcart" Utility =================== The new Irwin utility "/bin/mcart" has also been added to your system. It enables you to perform a number a useful functions on your Irwin tape drive. They are summarized below: h - This is a help command which allows you to view a list of available "mcart" commands. drive - This command provides information about drive type, controller type and unit select. info - This command provides information about cartridge state, format and write protect slider position. capacity - This command shows the total cartridge capacity in 512 byte blocks after formatting. kapacity - This command shows the total cartridge capacity in 1024 byte blocks after formatting. format - This command enables you to prepare unformatted tape cartridges for use with your Irwin tape drive. If you are using Accutrak pre-formatted tape cartridges, you will not need to use this command. (SEE NOTE BELOW) To use the "format" command, you need to make sure you are using either brand new, unformatted tapes, or bulk (magnetically) erased tapes. rewind - This command is used to rewind a tape after a backup or restore has been interrupted. Normally, the tape is rewound automatically. reten - This command is used to align a tape that has not been used for a long time, or has been exposed to temperature extremes. The command syntax is: mcart command For example, using the "rewind" option, the command would be: mcart rewind Notes ===== (1) Due to problems that arise while formatting tapes, SCO Technical Support is not able to fully support the "format" option. Keeping this in mind, for data integrity reasons, SCO highly recommends that pre-formatted tapes are used with Irwin tape drives. The format option is provided for use at your own risk! (2) The device you write to should be "/dev/rctmini". This device needs to be specified regardless of whether you are writing to the drive with "tar", "backup", "cpio", or some other utility. (3) Also, regardless of which utility you use to write to the drive, you MUST specify a blocking factor. The recommended blocking factor is "20". (4) If you choose to format your own tapes, remember that you MUST BULK ERASE the tape to be formatted. If you do not bulk erase your tape, you will not be able to format it. Please remember that the format option is *not* fully supported by SCO. It is provided for use at your own risk! (5) While installing your hardware, please refer to the hardware guide shipped with your Irwin tape backup unit. If you have questions regarding the installation of your Irwin hardware, contact Irwin Technical Support. If you have any further questions please contact your supplier. Yours sincerely, Technical Support Department ref: unx222 Addendum @@ CCCCCCCC Wait for Index ============== A condition exists in some Micro Channel computers which causes errors reading the first sector of each tape block. Included are the IBM PS2 models 50, 60, and 80, and the Olivetti P-500. If the following symptoms are experienced, after installing the MC driver in certain Micro Channel machines, the wait-for-index algorithm may need to be enabled: o On the first backup this message is seen: tape write error: Defect list has unrecoverable error o If "mcart format" gives the error: Formatting failed: Block 0 medium error : phase: CERTIFICATION, track: 0, cylinder: 0 o Extremely poor performance is experienced while listing the content of or restoring a previously written tape. To enable wait-for-index at the root prompt, the following commands may be used: # echo iropt=I >> /etc/default/mcconfig # reboot Using the wait-for-index algorithm has the following drawback: All user system task time processing is stopped until index polling is complete. This means the user will see sluggish system performance at certain times. Typically this means a 3 or 4 second dead period at tape track switch time. This may prove unacceptable in certain installations. Error Messages ============== Below is a list of the most common error messages included in the Irwin tape driver. Block not found =============== IE_BLKNOTFND(5) A block not found error occurs when the driver cannot locate a physical tape block during a read or write operation. Insure the tape head is clean (see the tape drive hardware manual for cleaning instructions). When this message is displayed during a data restore operation, try retensioning the tape, then repeat the restore operation. If this fails, try restoring the data using a different tape drive. When this message is displayed during data backup operation, try another tape. If your backup is successful on another tape, discard or bulk erase and reformat the original tape. Data CRC error ============= IE_CRC(12) ID CRC error ============ IE_IDCRC(13) An "ID CRC error", or "Data CRC error" message is displayed during a tape read operation when a tape block cannot be recovered by ECC. If this message appears, retension the tape and try again. If this fails, the data might be recovered by using a different tape drive. Causes of persistent CRC errors are: poor quality tapes, worn tape head, a defect in the drive's record circuitry, or an incompatible or otherwise defective data separator circuit on the controller. CRC errors might be stopped by using new tapes, or installing a data compensator circuit on the drive. Record not found ================ IE_RECNOTFND(14) A "Record not found" error occurs when some sector within a tape block cannot be located. This error typically occurs on a during a tape read operation when there are too many erroneous sectors to recover data using ECC. (See Data CRC Error). Drive not found =============== IE_DRVNOTFND(23) When /etc/mcdaemon is run (usually by an "rc" script) for the first time after system boot, a drive searching algorithm is executed by the driver. If this algorithm fails to detect the presence of a tape drive, " Drive not found" is displayed. Subsequently, the same message is displayed on any read or write access to an opened MC device file for which no drive is present. When this message appears, hardware should be checked. Servo failure ============= IE_SERVO(24) "Servo failure" is a tape formatting error message. The servo writing function is a part of the tape drive's firmware. The driver issues a command to the drive to servo write and awaits tape drive completion status. Usually the drive's LED will be flashing on servo writing failure. Try bulk erasing the cartridge and restart the format. If this fails, try another cartridge. Insufficient memory =================== IE_NOMEM(42) "Insufficient memory" messages are given when the driver fails to allocate at least three tape block buffers. Sufficient memory may be available when single user mode is entered immediately after system boot. Usually adding memory to the system will stop this message. Block 0 missing servo header IE_SERVOBLK0(43) Too many sequential missing servo headers IE_SERVOSEQ(44) Too many missing servo headers on track IE_SERVOTRK(45) Too many missing servo headers IE_SERVOTOT(46) Too many sequential bad blocks IE_CYLSEQ(47) Too many bad blocks on a track IE_CYLTRK(48) Too many bad blocks IE_CYLTOT(49) These messages occur during formatting. When formatting fails for one of these reasons, try bulk erasing the cartridge and reformatting. If this fails, try another cartridge. Block 0 medium error ==================== IE_BLK0DEFECT(52) A "Block 0 medium error" results when, during tape state determination, the first tape block has a medium error which is not recoverable by either redundant correction or ECC. Normally the cartridge should be reformatted or discarded. If the cartridge has a backup on it, try using a different drive to read the tape. If this message persists for multiple cartridges, the tape drive's read circuitry may be out of adjustment, or there may be an incompatibility between the floppy controller's data separator circuit and the tape drive. When running on a Micro Channel based machine the problem may be overcome. (See section od Wait for Index in this documentation.) Defect list has unrecoverable error =================================== IE_DLISTDEFECT(54) This message is displayed when both copies of the relocation table (kept in the second and third good tape blocks) have unrecoverable medium errors or are otherwise corrupt. When this message is displayed on the first backup after drive installation in a Micro Channel machine, the wait-for-index patch may be needed. Defect list corrupt =================== IE_DLISTCORRUPT(55) This message is displayed if an error is found in the primary defect list kept in block 0. Reformat the tape. Daemon not started ================== IE_NODAEMON(67) The tape driver uses a single daemon process to encode ECC during tape write operations and recover data with ECC during tape reads. When the "Daemon not started" message occurs, execute /etc/mcdaemon and retry the tape operation. If you have any further questions please contact your supplier. Yours sincerely, Technical Support Department Ref: unx222