84 IBM Tivoli Storage Manager Versions 5.4 and 5.5 Technical Guide
9. With the Tivoli Storage Manager server setup based on the commands above, data is
written to AD2TAPEPOOL only when the schedule COPYACTIVE2TAPE_BACKUPPOOL
is run or if the COPY ACTIVEDATA command is used. We chose this setup as it does not
require a mount point in the tape library during backup, which would be required if
synchronous write was configured. If you want to configure for synchronous write, update
the ACTIVEDATAPOOLS parameter for your primary storage pool to point to the
active-data pool, as shown in Example 7-19.
Example 7-19 Allow synchronous write to active-data pool
tsm: TSMAIX55>UPDATE STGPOOL BACKUPPOOL ACTIVEDATAPOOLS=ADPPOOL, AD2TAPEPOOL
ANR2202I Storage pool BACKUPPOOL updated.
Note that although we have configured two active-data pools for synchronous write from
BACKUPPOOL, since our two policy domains each specify a different active-data pool
destination, only one pool is used for synchronous write by each client.
You can now start including your active-data pool volumes as part of your physical or
electronic vaulting configuration. Doing this saves you storage and bandwidth by copying and
restoring only the active data.
If you are using active-data pools to replace a current copy storage pool, you could delete the
schedule to back up your primary pool, and update the primary pool so that no copy storage
pool is defined. Once all active data has been copied to the active-data pool you then could
delete the copy storage pool. Only perform these steps if you are replacing the copy storage
pool with the active-data pool. We generally recommend that both be maintained if having the
ability to restore inactive versions of backup data is required. Also, if the copy storage pool
contains archive or HSM files, do not delete the copy storage pool.
7.2.3 Restore client objects from an ADP
When restoring client data, the server selects the active version of a file from an active-data
pool when appropriate. The following restore order applies:
1. Active-data pool (FILE)
2. DISK (random)
3. FILE (primary or copyboy)
4. Active-data pool sequential onsite
5. Sequential onsite-volume (primary or copy)
a. mounted (idle)
6. Sequential offsite-volume (primary, copy or ADP)
Chapter 7. Active-data pools 85
Restore when there is no ADP
To simulate a restore without an ADP configured, we set the active-data pool’s volume to
UNAVAILABLE. We configured BACKUPPOOL to migrate to a tape primary storage pool,
and set the migration thresholds to 0 to empty BACKUPPOOL completely. We then started a
client restore. Since the data was not available in BACKUPPOOL, and the tape volume in the
active-data pool was UNAVAILABLE, the restore then mounted two tape volumes from the
tape primary pool, as shown in the activity log records in Example 7-20.
Example 7-20 Volume selection during restore where ADP is not available
ANR0406I Session 123 started for node ACTNODE1 (WinNT).
ANR1183I Initial determination of removable volumes required for a restore request
from session 123 is complete. Additional volumes may still be required.
ANR8337I LTO volume 936AAFL4 mounted in drive DRIVE1 (/dev/rmt0).
ANR0510I Session 123 opened input volume 936AAFL4.
ANR0514I Session 123 closed volume 936AAFL4.
ANR8468I LTO volume 936AAFL4 dismounted from drive DRIVE1 (/dev/rmt0) in library
ANR8337I LTO volume 937AAFL4 mounted in drive DRIVE1 (/dev/rmt0).
ANR0510I Session 123 opened input volume 937AAFL4.
ANR0514I Session 123 closed volume 937AAFL4.
The client restore statistics are shown in Example 7-21.
Example 7-21 Client restore, no ADP
tsm> res c:\demodata\ -repl=yes
ANS1247I Waiting for files from the server...
Total number of objects restored: 104
Total number of objects failed: 0
Total number of bytes transferred: 173.03 KB
Data transfer time: 78.70 sec
Network data transfer rate: 2.19 KB/sec
Aggregate data transfer rate: 1.38 KB/sec
Elapsed processing time: 00:02:04
Restore when an ADP exists and is available
We changed the access mode for the ADP volumes to read/write and performed the same
restore operation. The server activity log shows that this time, the data is restored from the
ADP, as shown in Example 7-22.
Example 7-22 Volume selection during ADP NQR restore
ANR0406I Session 125 started for node ACTNODE1 (WinNT).
ANR1183I Initial determination of removable volumes required for a restore
request from session 125 is complete. Additional volumes may still be required.
ANR8340I FILE volume /tsm/stg/activedata/0000002B.BFS mounted.
ANR0510I Session 125 opened input volume /tsm/stg/activedata/0000002B.BFS.
ANR0514I Session 125 closed volume /tsm/stg/activedata/0000002B.BFS.