Chi-Hsuan Huang created HDDS-15493:
--------------------------------------
Summary: Limit checkpoint-format parameterization in
TestOMRatisSnapshots to transfer tests
Key: HDDS-15493
URL: https://issues.apache.org/jira/browse/HDDS-15493
Project: Apache Ozone
Issue Type: Sub-task
Reporter: Chi-Hsuan Huang
HDDS\-14721 made the whole TestOMRatisSnapshots class parameterized \(via
@ParameterizedClass\) over ozone.om.db.checkpoint.use.inode.based.transfer
\(false/true\), doubling its runtime.
Measurements done under HDDS\-10310 show the parameter only affects the
checkpoint transfer path:
\- testInstallOldCheckpointFailure and testInstallCorruptedCheckpointFailure
never download a tarball \(they call installCheckpoint\(\) directly with a
locally prepared checkpoint\), so the parameter adds no coverage for them.
\- testInstallSnapshotWithClientWrite/ClientRead exercise the download, but
their v1/v2 format coverage duplicates TestOMBootstrap \(added by HDDS\-14721
specifically to verify both endpoints\); their unique value \(concurrent client
IO during install\) is format\-agnostic.
\- testInstallSnapshot is the only test covering the multi\-tarball path
\(ozone.om.ratis.snapshot.max.total.sst.size \+ SnapshotMaxSizeInjector\),
which TestOMBootstrap does not cover.
Proposal: move testInstallSnapshot \(plus the currently @Unhealthy
incremental\-snapshot tests and SnapshotMaxSizeInjector\) into a new
parameterized class \(e.g. TestOMRatisSnapshotTransfer\) that keeps running
under both checkpoint formats, and drop the class\-level parameterization from
the remaining tests \(they run once with the default, inode\-based transfer\).
Both formats stay covered on every transfer path, while the non\-transfer tests
stop paying the 2x cost; estimated saving is roughly 175s on top of HDDS\-10310.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
