[CASSANDRA-19345] Not using min/max column name meta data from CASSANDRA-8180 when the LIMIT clause is used

• Type: Bug
• Status: Open
• Priority: Normal
• Resolution: Unresolved
• Fix Version/s: 4.0.x, 4.1.x, 5.0.x, 5.x
• Component/s: Legacy/Local Write-Read Paths
• Bug Category: Degradation - Performance Bug/Regression
• Severity: Normal
• Complexity: Normal
• Discovered By: User Report
• Platform: All
• Assignee: Unassigned
• Reporter: Avneesh Pandey
• Created: 30/Jan/24 21:37
• Updated: 02/Feb/24 10:44

Description

CREATE TABLE tracking_details (
    tracker_public_id text,
    updated_at timeuuid,
    details text,
    PRIMARY KEY (tracker_public_id, updated_at)
) WITH CLUSTERING ORDER BY (updated_at DESC)

INSERT INTO tracking_details  (tracker_public_id, updated_at, details) VALUES('trk_e0544f077f7f4140bf1b961493fa99fe', now(), '{"details":[{...1-3KB }]}');

nodetool flush trackingdata tracking_details

INSERT INTO tracking_details  (tracker_public_id, updated_at, details) VALUES('trk_e0544f077f7f4140bf1b961493fa99fe', now(), '{"details":[{...1-3KB }]}');

INSERT INTO tracking_details  (tracker_public_id, updated_at, details) VALUES('trk_e0544f077f7f4140bf1b961493fa99fe', now(), '{"details":[{...1-3KB }]}');

nodetool flush trackingdata tracking_details

Compaction is disabled- 
 {'class': 'org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy', 'enabled': 'false', 'max_threshold': '32', 'min_threshold': '4'}

show version
[cqlsh 6.0.0 | Cassandra 4.0.5 | CQL spec 3.4.5 | Native protocol v5]

Each node is reading from both sstables - Merged data from memtables and 2 sstables. 
I thought this was fixed - https://issues.apache.org/jira/browse/CASSANDRA-8180_
But somehow  C* doesn't optimizes the read path

Dev Cluster Info - 3 Nodes, RF is 3 and CL is LOCAL_QUORUM

Tracing attached

Issue Links

• is related to: CASSANDRA-8180 Optimize disk seek using min/max column name meta data when the LIMIT clause is used