Device and volume error counters

When using the standard ROBAR rule file, ROBAR utilizes device and volume (cartridge) error counters as follows:

1. If an error occurs on a device (usually message EXC0858), ROBAR registers the error for the device in question. If a cartridge is mounted in the device at the time the error occurs, then ROBAR also registers an error for the cartridge.

2. When the error has been registered, the error handling procedure appropriate to the error is started (e.g. the cartridge is removed from the device).

3. ROBAR checks the device error counter for the device. If the defined limit (see the device_error_limit configuration parameter) has been reached, ROBAR begins with the relevant actions both for the job and for the device. (The actions are provided with default settings, but these can be changed to meet the requirements of your data center.)

   • Action key MRV26 in the ROBAR rule file message/message_resp defines the actions for the job. The following possibilities are available:

     MOVE	Move the job to another device (default).
     CANCEL	Cancel the job.
     IGNORE	Ignore the error (i.e. no action).

     The action defined with action key MRV26 is only started if the job affected by the error is still being processed. If, however, an error occurs while a cartridge is being removed, either no job is being processed on the device at that point in time or the job using the device has already changed and is not affected by the error.
     If MAREN has caused a MOUNT (message MARxxxx), the action is not started; the only reaction that is permitted is tsn.N

   • Action key MRV24 in the ROBAR rule file message/message_resp defines the actions for the device. The following possibilities are available:

     DETACH	Detach the device from the BS2000 system with /DETACH-DEVICE mn,FORCE=*YES (default).
     IGNORE	Ignore the error (i.e. no action).

     If you select the option IGNORE, the device continues to be available to the robot. Any new mount request issued for the cartridge in the course of this job is, however, rejected with the action key MRV027.

4. If an error counter exists for the cartridge (see the volume_error_limit configuration parameter), ROBAR checks it. If the defined limit has been reached, ROBAR starts the specified actions both for the job and for the cartridge (the actions are supplied with default settings, but you can adjust them to the specific requirements of your data center).

   • Action key MRV27 in the ROBAR rule file message/message_resp defines the actions for the job. The following possibilities are available:

     CANCEL	Cancel the job (default).
     IGNORE	Ignore the error (i.e. no action).

     The action defined with action key MRV27 is only started if the job affected by the error is still being processed. If, however, an error occurs while a cartridge is being removed, either no job is being processed on the device at that point in time or the job using the device has already changed and is not affected by the error.
     If MAREN has caused a MOUNT (message MARxxxx), the action is not started; the only reaction that is permitted is tsn.N

   • Action key MRV25 in the ROBAR rule file message/message_resp defines the actions for the cartridge. The following possibilities are available:

     EJECT	Place the cartridge in the output area (default). This option is not available in RRF CentricStor.
     IGNORE	Ignore the error (i.e. no action).

     If you select EJECT, the MTC error counter is reset as soon as the cartridge has been removed.

     If you select IGNORE, the device continues to be available to the robot. Any new mount request issued for the cartridge in the course of this job is, however, rejected with the action key MRV027.

Updating the MAREN catalog

When the error counter of a cartridge (see the volume_error_limit configuration parameter) reaches its maximum value, the ROBAR rule file causes the cartridge to be ejected into the input/output unit.

A notification is sent to ROBAR when this situation occurs, and the MAREN catalog is also updated.

Error counters and attached drives

If a device is attached to the system, ROBAR assumes by default that the device works free of errors and that it is empty (i.e. the archive system command ULK issues a message indicating that the device is free or a successful KEEP command is sent to the archive system). The device error counter is therefore reset in the following cases:

• If a device is empty and its status is “attached and assigned” or “detach pending“, as soon as a connection to a BS2000 system is established

• As soon as an empty device is attached in the BS2000 system

If the device error counter is not to be reset, resetting has to be deactivated by making the necessary modifications to the MRV28 action key in the ROBAR rule file
message/message_resp. The following settings are possible:

RESET	Reset the error counter (default).
DO NOTHING	Do not reset the error counter.

Device error counters can be reset manually in the ROBAR-SV Manager in the Diagnostics tab of the instance, see "Diagnostic information on the ROBAR-SV instance".