[
https://issues.apache.org/jira/browse/CASSANDRA-18062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17711861#comment-17711861
]
Andres de la Peña commented on CASSANDRA-18062:
-----------------------------------------------
I agree to merge this to the feature branch without fixing the flakies, so we
can parallelize work.
Most of the found flakies only fail once or twice, Thus, the mandatory minimum
number of repeated runs when merging CASSANDRA-16052 is likely to not hit all
of them. So let's keep around the above list of flakies, so we can give them
more CI rounds before merging the feature branch to trunk.
The CircleCI config to run these flakies can be created by removing [these
lines|https://github.com/apache/cassandra/blob/trunk/.circleci/generate.sh#L212-L218]
in {{generate.sh}} to disable the automatic detection of tests and manually
specifying the tests to be repeated:
{code:java}
.circleci/generate.sh -p \
-e REPEATED_UTESTS_COUNT=500 \
-e
REPEATED_UTESTS=org.apache.cassandra.index.sai.cql.DecimalLargeValueTest,org.apache.cassandra.index.sai.cql.MixedIndexImplementationsTest,org.apache.cassandra.index.sai.cql.StorageAttachedIndexDDLTest,org.apache.cassandra.index.sai.disk.SelectiveIntersectionTest,org.apache.cassandra.index.sai.functional.CompactionTest,org.apache.cassandra.index.sai.metrics.FinalSegmentFlushingFailureTest
\
-e REPEATED_JVM_DTESTS_COUNT=500 \
-e
REPEATED_JVM_DTESTS=org.apache.cassandra.distributed.test.sai.IndexAvailabilityTest
{code}
I have started such run
[here|https://app.circleci.com/pipelines/github/adelapena/cassandra/2843/workflows/87ec9c0e-9c87-4d97-ab8d-45beaad12088].
[The
config|https://github.com/adelapena/cassandra/commit/0f534d5880f9802c85b1bdced2916a6daf1e471d]
can be reused when trying to fix the flakies, possibly altering the number of
repetitions to whatever we feel comfortable with.
> On-disk string index with index building and on-disk query path
> ---------------------------------------------------------------
>
> Key: CASSANDRA-18062
> URL: https://issues.apache.org/jira/browse/CASSANDRA-18062
> Project: Cassandra
> Issue Type: New Feature
> Components: Feature/SAI
> Reporter: Mike Adamson
> Assignee: Mike Adamson
> Priority: Normal
> Labels: SAI
> Fix For: NA
>
>
> An on-disk index for string (literal) datatypes. This index is used for the
> following datatypes:
> * UTF8Type
> * AsciiType
> * CompositeType
> * Frozen types
> This includes the ability to write the index to disk at index creation, by
> specific index rebuild and during SSTable compaction.
> Also the ability to query the on-disk index and combine the results with
> those from the in-memory indexes created by CASSANDRA-18058.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
