Engine partitioning configuration

The process-instance partitioning and archiving configuration consolidated under flowx.data.partitioning.archiving. This page maps the old 5.1.x keys to the 5.9.x structure.

Who this affects

Only deployments that had data partitioning enabled on 5.1.x. If you never set FLOWX_DATA_PARTITIONING_ENABLED=true, there is nothing to migrate — the defaults are unchanged and partitioning stays off.

If you did use partitioning, the configuration shape changed between 5.1.x and 5.9.x. You must rename your settings before the upgrade; the old keys no longer bind.

---

What changed

In 5.1.x, partitioning settings and the archiving scheduler lived in two separate blocks — flowx.data.partitioning.* for partitioning and a standalone scheduler.dataPartitioning.* block for the archiving cron job.

In 5.9.x, everything is consolidated under a single flowx.data.partitioning block, with all archiving settings nested in an archiving sub-block. The standalone SCHEDULER_DATA_PARTITIONING_* variables no longer exist.

This is a hard cutover. 5.9.x does not read the old SCHEDULER_DATA_PARTITIONING_* keys or the old flat archiving variables. Settings left under the old names are silently ignored, falling back to defaults.

---

Configuration: before and after

5.1.x (old):

flowx:
  data:
    partitioning:
      enabled: true
      interval: MONTH
      retention-intervals: 3
      detached-partition-compression: OFF
      moved-data-batch-size: 5000
scheduler:
  dataPartitioning:
    enabled: true
    cronExpression: "0 0 1 * * ?"

5.9.x (new):

flowx:
  data:
    partitioning:
      enabled: true
      interval: MONTH
      archiving:
        enabled: true                  # was scheduler.dataPartitioning.enabled
        cronExpression: "0 0 1 * * ?"   # was scheduler.dataPartitioning.cronExpression
        retention-intervals: 3          # moved under archiving
        detached-partition-compression: OFF   # moved under archiving (Oracle only)
        moved-data-batch-size: 5000           # moved under archiving (PostgreSQL only)

---

Environment variable mapping

5.1.x variable	5.9.x variable
FLOWX_DATA_PARTITIONING_ENABLED	FLOWX_DATA_PARTITIONING_ENABLED (unchanged)
FLOWX_DATA_PARTITIONING_INTERVAL	FLOWX_DATA_PARTITIONING_INTERVAL (unchanged)
SCHEDULER_DATA_PARTITIONING_ENABLED	FLOWX_DATA_PARTITIONING_ARCHIVING_ENABLED
SCHEDULER_DATA_PARTITIONING_CRON_EXPRESSION	FLOWX_DATA_PARTITIONING_ARCHIVING_CRONEXPRESSION
FLOWX_DATA_PARTITIONING_RETENTIONINTERVALS	FLOWX_DATA_PARTITIONING_ARCHIVING_RETENTIONINTERVALS
FLOWX_DATA_PARTITIONING_DETACHEDPARTITIONCOMPRESSION	FLOWX_DATA_PARTITIONING_ARCHIVING_DETACHEDPARTITIONCOMPRESSION
FLOWX_DATA_PARTITIONING_MOVEDDATABATCHSIZE	FLOWX_DATA_PARTITIONING_ARCHIVING_MOVEDDATABATCHSIZE

---

Default behavior change

The archiving scheduler now defaults to off. In 5.1.x the scheduler defaulted to on (SCHEDULER_DATA_PARTITIONING_ENABLED=true); in 5.9.x archiving.enabled defaults to false. If you relied on the old default to run archiving, you must explicitly set FLOWX_DATA_PARTITIONING_ARCHIVING_ENABLED=true after the upgrade, or partitions will accumulate without being archived.

---

Migration steps

1. Move archiving settings under the archiving block

   Rename your retention, compression, and batch-size settings to the FLOWX_DATA_PARTITIONING_ARCHIVING_* form shown in the mapping table.

2. Replace the standalone scheduler variables

   Drop SCHEDULER_DATA_PARTITIONING_ENABLED and SCHEDULER_DATA_PARTITIONING_CRON_EXPRESSION. Set FLOWX_DATA_PARTITIONING_ARCHIVING_ENABLED=true and FLOWX_DATA_PARTITIONING_ARCHIVING_CRONEXPRESSION in their place.

3. Re-enable archiving explicitly

   Because the scheduler now defaults to off, confirm FLOWX_DATA_PARTITIONING_ARCHIVING_ENABLED=true is set if you want archiving to run.

4. Redeploy and verify

   Redeploy process-engine and confirm the archiving cron job runs on schedule. Check the DATA_PARTITIONING_LOG table for new archiving entries.

---

Integration Designer

5.9.x also adds partitioning and archiving for workflow instances in Integration Designer, using the same flowx.data.partitioning structure. It is opt-in and defaults to off, so it is not a required migration step — enable it only if you want to partition workflow instance data. Note that the Integration Designer block does not include the Oracle compression or PostgreSQL batch-size settings.

---

Rollback

The configuration rename on this page is reversible. It is configuration-only, and the database schema is forward-compatible, so reverting to the old block names restores the previous behavior without a schema change.

This does not apply to the native PostgreSQL partitioning conversion available starting with the 5.9.5 patch (and in 5.12). That one does change the schema and is one way. See Native PostgreSQL partitioning.

---

Related resources

Partitioning & archiving setup

Full partitioning and archiving configuration reference.

Prerequisites & upgrade path

Inventory to gather before the upgrade, including your current partitioning config.