[
https://issues.apache.org/jira/browse/CASSANDRA-20954?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18029894#comment-18029894
]
Caleb Rackliffe commented on CASSANDRA-20954:
---------------------------------------------
I think {{MultiNodeTableWalkWithMutationTrackingTest}} reproduces this with
{{builder.withSeed(3448512636059630802L).withExamples(1);}}
{noformat}
Caused by: java.lang.AssertionError: Unexpected results for query: SELECT *
FROM ks1.tbl WHERE v0 >= [{0}, {0, 514}, {-1715, 3, 1215135}] PER PARTITION
LIMIT 140 LIMIT 10 ALLOW FILTERING
Caused by: java.lang.AssertionError: Missing rows:
pk0 | ck0 | v0 | v1
'2025-01-28T20:50:22.330Z' | -60 | [{7948, 43651062, 65346426}, {0}, {7}] |
{f0: {false: true, true: true}}
'2025-01-28T20:50:22.330Z' | -34 | [{7948, 43651062, 65346426}, {0}, {7}] |
{f0: {false: true, true: true}}
'2025-01-28T20:50:22.330Z' | 21 | [{15}, {1176775}, {5416, 8072}] |
null
Expected:
pk0 | ck0 | v0
| v1
'2022-12-30T09:50:17.410Z' | -112 | [{0, 51375431, 3145580081}, {-2630377, 24},
{-20506, -1, 2}] | null
'1996-03-11T12:10:30.805Z' | -83 | [{111}, {-1588}, {-73, 0, 5105}]
| {f0: {false: false}}
'1996-03-11T12:10:30.805Z' | -77 | [{0, 87, 60579}, {319196}, {0, 13}]
| {f0: {false: false, true: false}}
'1996-03-11T12:10:30.805Z' | -28 | [{0}, {0, 514}, {-1715, 3, 1215135}]
| null
'2025-01-28T20:50:22.330Z' | -60 | [{7948, 43651062, 65346426}, {0}, {7}]
| {f0: {false: true, true: true}}
'2025-01-28T20:50:22.330Z' | -34 | [{7948, 43651062, 65346426}, {0}, {7}]
| {f0: {false: true, true: true}}
'2025-01-28T20:50:22.330Z' | 21 | [{15}, {1176775}, {5416, 8072}]
| null
{noformat}
> (CEP-45) Fix SRP in range reads with PER PARTITION LIMIT set, simplify the
> read path
> ------------------------------------------------------------------------------------
>
> Key: CASSANDRA-20954
> URL: https://issues.apache.org/jira/browse/CASSANDRA-20954
> Project: Apache Cassandra
> Issue Type: Bug
> Components: Consistency/Coordination
> Reporter: Aleksey Yeschenko
> Assignee: Aleksey Yeschenko
> Priority: Normal
>
> There is a slight issue with SRP in the current mutation tracking range read
> implementation - if an augmenting mutation partially shadows a local LSM
> read, making a partition with PER PARTITION LIMIT set short, then SRP won't
> be invoked for that partition.
> While at it, we might simplify the read state machine class hierarchy and
> make obtaining follow-up reads for SRP a part of the simplified state machine
> rather than have a mixture of state machine transitions and
> callbacks/promises, giving it a more linear structure.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
