We aim to reduce the thin_check's execution time to improve the responsiveness of LVM thin-pool activation and error checking. The v1.0 release has already brought significant performance improvements, reducing the time to examine a 16 GB metadata on an SSD from over 5 minutes in v0.9.0 to just 30 to 40 seconds. However, performance bottlenecks remain, and v1.0 does not scale well on multi-core systems. For example, adding more than two worker threads yields little to no additional performance gain, even though the current configuration uses four worker threads.

Goal

• As a user, I want thin_check to run twice as fast so that I can activate thin-pools more quickly.

Acceptance criteria

The POC has shown that a runtime reduction of approximately 50% is achievable. For example, on a laptop system, checking the 16 GB metadata from RHBZ 1763895 should complete within 15 seconds, compared to 32 seconds in the current release.