Autochanger Operations

This section explains how to resolve problems encountered with the use of an autochanger with Backup.

Testing the Device Driver Installation

After you install the Backup device driver software, use the lusdebug program to verify the server connection and the jbexercise program to test the autochanger. Use the value of the control port assigned to your autochanger (for example, scsidev@0.6.0) for control-port in the following commands:

# lusdebug control-port 0
# jbexercise -c control-port -m model

If these commands fail, or if you receive error messages, see the following sections for information on the possible cause and solution.

The lusdebug Command Fails

If the lusdebug command fails, review these suggestions to identify the potential problems and their solutions:

• Issue the sjiinq command as root, and provide the control-port as an argument. You should receive a message similar to the following:

scsidev@0.6.0:<EXABYTE EXB-10i EXB-10i >

Verify that the information supplied by the message is correct.

If the vendor and model names are incorrect, you supplied the wrong SCSI ID as the device ID during the driver installation. The installation script asks for the SCSI ID of the robot mechanism, not the tape drive.

• Deinstall the device driver and then reinstall it, and supply the correct address for the autochanger (robotic arm). Make sure that each device on the SCSI bus has a different SCSI ID address.

• Inspect the following items to verify that the autochanger is properly connected:

  • Make sure all the connectors on the SCSI bus are firmly connected.

  • Make sure none of the SCSI cables are defective.

  • Verify that the SCSI bus is properly terminated and is within the length specified by ANSI SCSI-II specifications (ANSI X3.131-1994).

    Both ends of the SCSI bus must be terminated with the appropriate resistors to be properly terminated. Single-ended SCSI buses are 220 ohms to +5 VDC, 330 ohms to ground. Differential terminators have a 122-ohm characteristic impedance (-5 VDC to +5 VDC). The ends of the SCSI bus are considered to be the last SCSI device at either end of the bus, where both peripheral devices and systems are considered as peer SCSI devices.

    Additional termination (termination placed on devices not at either end of the SCSI bus) is ill-advised. Additional termination causes the hardware bus drivers on each device on the bus to have to work harder (for example, out of the range of their nominal specification) to affect signal transitions. As a result, they may not be able to meet the timing requirements for some signal transitions.

    SCSI bus length limitations affect the quality of the signals; thus, the likelihood of transmission errors on the bus. For Single-ended SCSI buses (the most prevalent), the length is 6 meters, unless FAST SCSI devices are attached and in use, in which case the length limit is 3 meters. This length includes the length of the bus as it is within a device as well as the length of external cables. A reasonable rule of thumb for internal length is to assume 1 meter of internal bus length for the workstation chassis and about 0.25 meters per device for external peripheral boxes.

    Differential option SCSI buses can be much longer (due to the electrical differences from Single-Ended). Allow for a maximum of 25 meters. Never mix Differential and Single-Ended devices.

• Check to see whether an old autochanger driver is still installed. This can be the AAP driver shipped with earlier versions of Backup, or release 1.1 or earlier of the Parity driver, which only supported SCSI bus 0.

• Check the SCSI IDs on all devices attached to the same bus; make sure that none are the same. If two devices have the same target ID, you may experience the following symptoms: SCSI bus reset errors appear in system log files, the machine does not boot, and the probe-scsi boot prompt command on SPARC systems hangs.

• If the sensor that verifies whether the tape drive door is open is out of place, follow the instructions provided with your autochanger hardware to determine the problem, or contact your hardware vendor.

• If the autochanger is in sequential mode, change the setting to random mode.

If none of these suggestions resolve the problem, contact SunSoft Technical Support. You need to provide the information described in "Information to Gather Before You Call Technical Support ", and the captured output of the sjiinq, and sjirjc programs. See Appendix B, Command Line Reference Utilities for information on the jbexercise, sjiinq, and sjirjc programs, or refer to the associated man pages for each program.

The jbexercise Command Fails

If the jbexercise command fails, review the following list of suggestions to identify potential problems and their solutions:

• The jbexercise program prompts you for a no-rewind device name (for example, on Solaris, /dev/rmt/0mbn). Verify that you have supplied the correct device pathname for the tape drive. The device name must belong to a tape drive in the autochanger, not the autochanger itself. If you receive the following error message, you did not enter a no-rewind device name:

  device not ready

• Make sure that the tape drive for which you enter the pathname works. Insert a volume into the drive and perform the following tests:

  • Use the tar command to copy a small file to the volume.

  • Verify more extensive operations by issuing the tapeexercise command.

  If these tests fail, the tape drive is not functioning. Contact your hardware vendor for further information on how to configure your tape drive to work with your system.

If none of these suggestions resolve the problem, contact SunSoft Technical Support. You need to provide the information described in "Information to Gather Before You Call Technical Support " and the captured output of the jbexercise, sjiinq, and sjirjc programs. See Appendix B, Command Line Reference Utilities for information on the jbexercise, sjiinq, and sjirjc programs, or refer to the associated man pages for each program.

Autodetected SCSI Jukebox Option Causes Server to Hang

If you install an Autodetected SCSI jukebox using jb_config and the server hangs, the following workaround is recommended:

1. Select the jb_config option that installs an SJI jukebox. A list of jukeboxes is displayed.

2. Enter the number that corresponds to the type of jukebox you are installing.

3. Proceed with jb_config until you receive the following message:

   Jukebox has been added successfully.

Autochanger Inventory Problems

The autochanger inventory becomes outdated, which means that Backup cannot use the autochanger, if any of the following situations occur:

• The media is manually ejected from the autochanger drive.

• The media is removed from the autochanger.

• The autochanger door is opened.

To make the autochanger usable again, perform the following steps:

1. Verify that the media cartridge is correctly installed in the autochanger and that the autochanger door is closed.

2. Become root on the Backup server.

3. Reset the autochanger.

   # nsrjb -Hv

4. Perform an inventory.

   # nsrjb -Iv

   After the inventory operation is finished, Backup can once again use the autochanger.

   For complete information on the use of the nsrjb command, refer to the nsrjb(8) man page or see Chapter 7, Autochanger Module.

Destination Component Full Messages

The message "Destination component full" usually is the result of a manual operation performed on the autochanger, for example, physically unloading the tape drive by means of the buttons on the autochanger rather than using Backup to unmount the volume. This operation causes Backup to lose track of the status of the media in the autochanger.

To resolve the problem, use Backup command nsrjb -H to reset the autochanger.

Tapes Are Not Filled to Capacity

You might encounter situations where Backup does not fill tapes to capacity. For example, a tape with an advertised capacity of 4000 MB can be marked full by Backup after only 3000 MB of data have been written to it.

To enable Backup to use the tape capacity to its fullest, select the highest density device driver appropriate for your device. When a tape is labeled, Backup writes to it at the highest density supported by your device.

There are several reasons for situations in which Backup appears to fill tapes prematurely:

• Write errors occur during a backup.

  Most tape drives try to read after a write operation to verify that the tape was written correctly, and retry if it was not. A write error indicates either end of tape or a read error. At any tape error, Backup marks the tape full.

  To prevent tape write errors, clean your tape drive regularly and use only data-quality tapes. If cleaning the drive does not seem to help, make sure that the device driver is properly configured, any necessary switch settings on the tape drive are set to the manufacturer's specifications, all cabling is secure, and other potential SCSI problems have been addressed.

• Backup filemarks take up space on the tape.

  Backup periodically writes filemarks to facilitate rapid recovery of data. These filemarks consume varying amounts of tape depending on the type of tape drive on some drives, filemarks can consume several MB. The number of filemarks Backup writes to tape is a function of how many save sets are on the tape. Many small save sets require more filemarks than a few larger ones.

• Tape capacities vary from tape to tape.

  Tape capacities are not constant from tape to tape. Two apparently identical tapes from the same vendor can vary significantly in capacities. This can cause problems if you copy one very full tape to another, especially if the destination tape holds less data than the source tape.

• Data compression affects the tape capacity.

  If you use compression on your tape drive, you cannot predict the effect on tape capacities. A compressing drive can provide twice the capacity of a non-compressing drive. It could be far less or far more, depending on the kind of data being backed up. For example, if a noncompressing drive writes 2 GB of data to a specific tape, the compressing drive could write 10 GB, 2 GB, 5 GB, or some other unpredictable amount of data.

• Length of tape.

  Be sure to verify tape lengths. A 120-meter DAT tape holds more data than a 90-meter DAT tape, and without examining the printed information on the tape cassette carefully, the two tapes can appear identical.

  For Solaris, if your tape devices are not directly supported by Sun Microsystems, you will need to recreate your entries in the st.conf file. If you need assistance with this, contact Sun Technical Support.

Server Cannot Access Autochanger Control Port

The control port controls the autochanger loading mechanism. Your autochanger's hardware installation manual should have instructions on how to verify whether the control port is properly connected. If you cannot determine whether the control port is working, contact the autochanger vendor for assistance.