Integrating SAP R/3 and Data Protector

 

Troubleshooting

 

Util_File_Online SAP backup fails with “semop() error”

Problem

When the util_file_online option is used with BRBACKUP (for

 

example, if you select the Brbackup_Util_File_Online template), the

 

tablespaces are switched into/from backup mode individually. As there

 

can be only one process communicating with BRBACKUP, several

 

sapback processes are using a semaphore to synchronize their

 

interaction with BRBACKUP.

 

The number of sapback processes is calculated as the sum of

 

concurrencies of all devices used for backup. With a large number of

 

sapback processes, the maximum number of processes that can have

 

undo operations pending on any given IPC semaphore on the system may

 

be exceeded. In such case, several sapback agents will fail with the

 

following error:

 

[28] No space left on device.

Action

Perform any of the following actions to resolve the problem:

 

• Reduce the number of backup devices or their concurrency.

 

• Increase the value of the semmnu kernel parameter. After you increase

 

the value, rebuild the kernel and reboot the system.

 

Restore Problems

 

At this stage, you should have performed all the verification steps

 

described in the previous sections. After this, proceed as follows:

 

1. Verify a user for the restore:

 

Verify that user specified for the restore session is the user of backup

 

session and that he/she belongs to the Data Protector operator or

 

admin group.

 

See “Configuring an SAP R/3 User in Data Protector (UNIX Systems

 

Only)” on page 165

 

2. Verify that a backup object exists on the backup media and in
the IDB:

This can be done by executing the command

/opt/omni/bin/omnidb -SAP "<object_name>" -session "<Session_ID>" -media (HP-UX and Solaris systems) or

/usr/omni/bin/omnidb -SAP "<object_name>" -session

Chapter 2

235