cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alex Petrov (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-12652) Failure in SASIIndexTest.testStaticIndex-compression
Date Mon, 07 Nov 2016 15:59:58 GMT


Alex Petrov commented on CASSANDRA-12652:

My assumption was as follows: since this is a "compression" run, there are some sstable leftovers
from the previous (non-compression) run. One run creates 2 sstables, second run creates 2
sstables. Since “low” threshold of STCS is 4, we trigger a compaction. After comaction,
index rebuild is asynchronous, so index might be empty, since there are no sasi memtables…
So there’re three scenarios: one when we’ve flushed the index, one where we haven’t
and one where we're still reading the old sstables. One of them (unflushed index) would fail.

I've tried running tests, where I would insert same data 2 more times (to create a total of
4 sstables) and "emulate" the long flush for the static cf (by adding a sleep when creating
a 5th index for the {{static_sasi_test_cf}}). The test results are as expected:

Expected :1
Actual   :0

Having that said, we might want to create another issue that would take care of making sure
that SASI indexes are served correctly for the compacted tables..

> Failure in SASIIndexTest.testStaticIndex-compression
> ----------------------------------------------------
>                 Key: CASSANDRA-12652
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Testing
>            Reporter: Joel Knighton
>            Assignee: Alex Petrov
>             Fix For: 3.x, 4.x
> Stacktrace:
> {code}
> junit.framework.AssertionFailedError: expected:<1> but was:<0>
> 	at org.apache.cassandra.index.sasi.SASIIndexTest.testStaticIndex(
> 	at org.apache.cassandra.index.sasi.SASIIndexTest.testStaticIndex(
> {code}
> Example failure:

This message was sent by Atlassian JIRA

View raw message