whpg-diskquota 2.4.0 release notes

Release date: September 20, 2026

whpg-diskquota 2.4.0 includes new features, improvements, a behavior change, and bug fixes.

Features

• Set a database quota: Limit the total disk space a single database can use. See Setting a database quota.
• Set a cluster quota: Limit the combined disk space used by every monitored database in the cluster. See Setting a cluster quota.
• View database and cluster quotas: Display the new quotas alongside current usage. See Displaying disk quotas and disk usage.
• Downgrade to 2.3: prepare_downgrade() prepares a database for downgrading to 2.3. See Downgrading from 2.4.
• ppc64le support: whpg-diskquota 2.4.0 is the first release built for the ppc64le architecture. See Supported platforms.

Improvements

• A paused database keeps reporting accurate usage, so it stays correctly counted toward the cluster quota. See Pausing and resuming quota enforcement.
• A monitored database that's dropped without DROP EXTENSION, which can happen in dynamic mode when no worker is connected to it, is now cleaned up automatically and stops counting toward the cluster quota.
• set_cluster_quota() now warns about databases it can't count towards the cluster quota, whether they're missing the extension or not yet initialized. See Setting a cluster quota.

Behavior changes

• Pausing no longer reduces load on your cluster. A paused database's worker still runs its full size refresh every diskquota.naptime. See Pausing and resuming quota enforcement.

Bug fixes

• Deleting a quota could leave writes blocked indefinitely when diskquota.hard_limit was off on the coordinator but still on at the segments, for example after a partial configuration reload, or when sending the rejectmap to the segments failed. whpg-diskquota now always sends quota removals to the segments, and retries a failed send. This bug affected every quota type, not only the new database and cluster quotas.
• Running DROP EXTENSION diskquota after the launcher restarted could crash the launcher.