Compression

Add the tables and indexes that you choose, and Scops moves them to the compression level that you set, inside the window of each database.

The Compression card of LEDGER, with the move that runs now, the GB saved per day and the Objects tab

Before you begin

  • View Compression opens the page. Manage Compression changes databases, objects and profiles. Run Compression starts, stops, discovers, recompresses, retries and clears.
  • The monitoring account needs the Compression grants, and ALTER SYSTEM to stop a move. See Feature grants.
  • The tablespace needs free space for a copy of each segment that moves.

Scops does not check your Oracle licenses or storage, so you confirm the Advanced Compression license for each database that you add. Advanced Row (OLTP), index and LOB compression need the Advanced Compression option, and an online move needs Enterprise Edition. The Query and Archive levels need Exadata or ZFS storage.

Add a database

  1. Open Add-ons › Compression and click Add Database.

  2. Select the database.

  3. Check the sections on the left. See Settings.

  4. In the footer, tick I confirm this database is licensed for Oracle Advanced Compression and I take responsibility for compressing it. You must tick it before you click Add Database. Scops records your user name and the time.

  5. Click Add Database.

  6. Add the objects. Scops compresses only the objects on the Objects list.

  7. If the badge next to the title shows Stopped, click Start Job.

Add objects

Add Objects for LEDGER, with three tables and one index ticked in schema FIN_APP
The ticks stay when you change the schema.
  1. On the database, open Objects and click Add Objects.

  2. Select a Schema and tick the tables and indexes.

  3. Click Continue.

  4. Switch each pass on or off, such as Compress tables or Compress LOBs, and select its profile.

    The Settings step of Add Objects, with the passes and profiles for three tables and one index
  5. Click the Add button, for example Add 4 Objects.

The next discovery queues the work. Remove on an object stops its work and keeps its history.

Force Recompress

Force Recompress moves listed objects again: to decompress them, to go to a lower level, or to rewrite segments that DML left partly uncompressed.

  1. On the database, open Objects.

  2. Move the pointer over the row of the object, then click the Settings icon.

  3. Under Now, switch on Force Recompress.

  4. Click Force Recompress. If you changed a setting, the button is Save and Force Recompress.

For many objects, click Select, tick the objects, click Force Recompress and click it again in the message.

Scops queues each segment one time, with no size floor, after the work that saves more space. The moves run in the window of the database.

To recompress objects that you add, switch on Force Recompress in the Settings step of Add Objects.

The Now group of Add Objects, with Force Recompress switched on

Set the window

  1. Click Settings on the database and open Schedule.

  2. Set the Days, Starts at and Ends at, in whole hours on the clock of the Scops server.

  3. Click Save.

Moves start only inside the window. At the end of the window, a running move goes back to the queue. Without a window, moves run at any time.

Compress partitions by age

The Age Rules section of LEDGER with two rules
Partitions older than 90 days go to Query High, older than a year to Archive High.
  1. Click Settings on the database and open Age Rules.

  2. Switch on Compress partitions by age.

  3. Click Add Rule. Set the Name, Partitions older than and Compress to.

  4. Click Add Rule, then Save.

Scops counts the age of a range partition from its upper bound, and the oldest rule that matches wins. Scops does not create Oracle ILM policies.

Update or clear the queue

  • Every hour, also while the job is stopped, discovery queues the segments that are below their level. It never lowers a level. To run it now, click Discover or Discover All.
  • To delete the queued tasks of a database, click Clear Queued on Queue. The next discovery queues them again.

Retry or clear failed tasks

The History tab of LEDGER, with one finished task opened on its sizes and statement
  1. On the database, open History.

  2. Open a failed row to see the statement and the error. Correct the cause.

  3. Click Retry Failed to queue the failed tasks again, or Clear Failed to delete them.

What happenedResult
You stopped the move, the window closed, or other DDL broke a parallel move (ORA-12842).Back in the queue. Not a failure.
The table was busy: ORA-00054, ORA-00060, ORA-04021 or ORA-30006.Failed. Scops queues it again each hour, up to 3 failures.
Any other error, such as no space, a missing grant, or a move longer than 4 hours.Failed. It waits for you.

Skipped means that nothing ran: the object was dropped, was at its level, left the list, or is a unique index on one column, which Oracle compresses only at Advanced High.

Stop and start

To stopDo this
All databasesClick Stop Job. The job starts again with Start Job, or when the next window opens.
One databaseClick Stop on the database. Start starts again.
Some objectsOn Objects, click Select, tick the objects and click Stop.
One database, for a long timeIn Settings, switch off Compression on and click Save. The settings stay.

Scops kills its sessions at once, and Oracle rolls back the move. The work goes back to the queue.

Without the ALTER SYSTEM grant, Scops cannot kill a move. The move continues to the end, but the job shows stopped.

Remove a database

In Settings, click Remove This Database. Scops deletes the objects and the task history of the database.

What you see

All Databases in Compression, with the tiles and the list of four databases

On All Databases, click a tile to see its tasks.

On Objects, Waiting · for discovery means that nothing is queued yet. Nothing runs means that all passes of the object are off. Too small means that the object is under the size floor of its profile: lower the floor or use Force Recompress.

The Queue tab of LEDGER with 41 steps in seven rows
Queue has a row for each table and kind of work, in run order. Uncompressed and larger segments go first.

Settings

The Tables and LOBs settings of LEDGER on the Warehouse HCC profile

Each section is a profile that databases and objects can share. Save changes the profile for all its users. To change one database only, click Save as New first.

SettingWhat it doesDefault
Compress tables and partitionsThe table pass.On
Compress toAdvanced Row (OLTP), a Query or Archive level, or None — decompress. A lower level needs Force Recompress.Advanced Row
During the moveOnline — application keeps writing, or Offline — exclusive lock.Online
Applies toTables, Partitions, Subpartitions, each with a size floor.Tables and Partitions; 50, 10, 1 GB
Parallel degreeParallel servers for a move, 1 to 64. The DEGREE of the table does not change.4
Moves at onceSegments that move at the same time.2
Move partitions of one table in parallelMoves the partitions of a table at the same time. This can cause library cache lock waits.Off
Rebuild indexes in the same moveAdds UPDATE INDEXES to an offline move.On
Compress LOB columnsCompresses SecureFile LOBs in the move of the table.Off; Medium, 1 GB
Refresh statistics after each moveGathers the statistics of the moved segment.Off
Compress indexesThe index pass: Advanced Low, Advanced High or None — decompress.Off; Advanced Low; 10, 5, 1 GB; degree 4, 2 at once
Look for wasteful compressed segmentsEvery four hours, measures the segments where DML touched 20% of the rows. Scops queues a move that frees 1 GB and a tenth of the segment.Off

At most 256 databases move segments at the same time: At once of the Compression job in Job Settings.

Updated November 3, 2026