Troubleshooting Shared Drive Problems

The following is a list of common potential problems when configuring shared drives. Also refer to the SSO chapter of the NetBackup Media Manager system administrator's guide.

• Using incompatible or outdated firmware or drivers in a hub, switch, HBA, or bridge.
• Did not set the JNI HBA fail-over value to a value of zero to avoid I/O hangs (this is a bridge/HBA vendor fix).
• Using a HBA with SCSI-3 protocol, and the HBA is not compatible with the operating system drivers.
• Using cluster configurations when they were not supported.
• Using vendor peripherals that only work on a fibre-channel arbitrated loop.
• Did not verify that SSO has been enabled on each server (you enable SSO using the Shared Drive license key).
• Did not verify that SSO has been installed correctly. You can check keys by using the license key GUI available from the NetBackup Help menu on Windows and UNIX servers.
• Did not configure all of SSO from the master server. All configuration should be done from the master server, not from a media server (or SAN media server).
• Did not configure the same robotic path on every host. Remember that except for ACS and TLM robot types, only one host controls the robot.
• When using the Device Configuration wizard, did not select the appropriate device hosts, including the host with robotic control.
• Created inconsistent configurations by using tpconfig to configure SSO rather using than the configuration wizards. These wizards have the added benefit of coordinating configurations across all hosts that are sharing the drives.
• Drives and robots that are connected by fibre channel cause increased complexity in a Media Manager device configuration. On some operating systems, the use of SCSI-to-fibre bridges may result in inconsistencies in the device paths when rebooting the host. After a reboot of the host, the device configuration should be verified.
• Using a name that is not consistent across all systems sharing drives.
• Did not test the drive paths on every media server.
• Did not define NetBackup storage units for each media server.
• Interrupting the data path while backup data is being transferred will cause the NetBackup job to fail. It can fail with a media write error or it may hang and have to be terminated manually.
• Did not use Berkeley-style close on the tape path (UNIX servers only)
• See the Sun chapter of the NetBackup Media Manager device configuration guide for more information on the following configuration tasks.

Forgot to add tape configuration list entries in /kernel/drv/st.conf (if needed).

Did not define configuration entries for expanded targets and LUNs in sg.links and sg.conf files. If you see problems with the entries in the /etc/devlink.tab file (created from sg.links). Check the following:

• The first entry uses hexadecimal notation for the target and LUN. The second entry uses decimal notation for the target and LUN.
• Use a single tab character between the entries, not a space or a space and a tab.

Did not configure the operating system to force load the sg/st/fcaw drivers.