This page summarizes the expected behavior differences between V1 and V2 volumes and provides a feature support matrix for the Longhorn v1.12.0 documentation set.

Expected Behavioral Differences

Snapshot Deletion Behavior

When deleting the snapshot that is the direct parent of volume-head, V1 and V2 volumes behave differently.

Snapshot Purge Behavior

Snapshot purge behavior also differs between V1 and V2 volumes.

In V1, purge is the operation that coalesces snapshots previously marked as removed. It runs separately from snapshot deletion and is subject to a per-node concurrency limit (Snapshot Heavy Task Concurrent Limit). As a result, removed snapshots may remain visible in the snapshot chain or API response until purge completes.

In V2, purge runs immediately and removes eligible system-created snapshots (those not created by the user) from the replica chain in one operation. The underlying storage layer performs the merge during the purge.

Revision Counter Behavior

Revision counter behavior differs between V1 and V2 volumes.

In V1, Longhorn supports revision counters for tracking replica updates. This mechanism can be used during startup and auto-salvage to help identify the replica with the latest update. For more information, see Revision Counter.

In V2, revision counters are not supported. V2 volumes do not maintain revision-counter-based replica tracking, and V1-specific revision counter settings do not apply.

Feature Support Matrix