HP XP7 Replication Manager Application Agent CLI User Guide (TK981-96015, May 2014)

Procedures for restoration and recovery are as follows after failure occurs at point D.
1. Restore the data in the SQL Server database backed up at point A.
The database is recovered to point A. For details about how to restore the SQL Server database
backed up at point A, see “Restoring SQL Server databases” on page 315. Also, for the order to
apply the transaction logs, see “Precautions regarding application of transaction logs” on page 285.
2. Perform recovery by applying the transaction logs backed up at point B and point C.
When you perform recovery by applying the transaction logs backed up at point B (point C), the
database reverts to that at point B (point C).
For details about how to perform recovery by using the drmsqlrecovertool dialog box and
applying the transaction logs, see “Using the drmsqlrecovertool dialog box to recover an SQL
Server database” on page 315.
When you apply multiple transaction logs for the drmsqlrecover command, you can apply
multiple transaction logs at one time by using the transaction log batch file. For details about
how to perform recovery using a transaction log list file, see “Using the drmsqlrecover command
to recover an SQL Server database (using a transaction log list file)” on page 316.
Backing up an SQL Server database
Assuming that the database is in the state as that at "point A", obtain a backup of the SQL Server
database.
Resynchronizing copy groups
When usual operation is in the split state, the copy group is first resynchronized, after which the data
is backed up.
When usual operation is in the pair status, it is unnecessary to resynchronize the copy group before
backing up the data. After the data is backed up to tape, resynchronize the copy group to restore
data to the initial state.
To resynchronize the copy group:
1. For multi-generation operations, check the copy group to be used next.
To check the copy group to be used next, execute the drmsqlcat command on the backup
server.
If the copy group is not locked, the least recently backed up copy group is overwritten. (You can
use the drmcgctl command to check the lock status of a copy group.)
2. Clear the cache of the secondary volume.
Before performing backup, clear the system cache of the backup server.
To clear the system cache, execute EX_DRM_CACHE_PURGE on the backup server at the local
site to mount, and then unmount the secondary volume.
BKServer > EX_DRM_CACHE_PURGE Operation_A -cg_file C:\temp\CGLIST.txt
In this example, the caches of the secondary volumes belonging to copy groups with names
VG01,vol11 and VG01,vol12 are cleared. To specify multiple copy group names, you can
create a copy-group list file ahead of time, and simply specify the file. In this example, copy
groups VG01,vol11 and VG01,vol12 are defined in the copy-group list file CGLIST.txt.
Application Agent CLI User Guide 313