[
https://issues.apache.org/jira/browse/CASSANDRA-13395?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15950928#comment-15950928
]
Benjamin Lerer commented on CASSANDRA-13395:
--------------------------------------------
I discussed offline with [~slebresne] and the internal iterators do not accept
empty rows for performance reasons.
As we know that except for indexes the deleted cells are caused by the
compaction of expired row marker we can avoid the empty row problem by treating
those rows as the expired ones. The only information missing being the original
TTL we can simply replace that one by a fake one.
I pushed an initial version of the patch
[here|https://github.com/apache/cassandra/compare/trunk...blerer:13395-3.0].
> Expired rows without regular column data can crash upgradesstables
> ------------------------------------------------------------------
>
> Key: CASSANDRA-13395
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13395
> Project: Cassandra
> Issue Type: Bug
> Reporter: Benjamin Lerer
> Assignee: Benjamin Lerer
>
> In {{2.x}} if an expired row is compacted its row marker will be converted
> into a {{DeletedCell}}. In {{3.0}}, when the row is read by {{LegacyLayout}}
> it will be converted in a row without {{PrimaryKeyLivenessInfo}}. If the row
> does not contains any data for the regular columns, or if the table simply
> has no regular columns it will then be considered as {{empty}}. Which will
> crash {{upgradesstables}} with the following error:
> {code}
> java.lang.AssertionError
> at org.apache.cassandra.db.rows.Rows.collectStats(Rows.java:70)
> at
> org.apache.cassandra.io.sstable.format.big.BigTableWriter$StatsCollector.applyToRow(BigTableWriter.java:207)
> at
> org.apache.cassandra.db.transform.BaseRows.applyOne(BaseRows.java:116)
> at org.apache.cassandra.db.transform.BaseRows.add(BaseRows.java:107)
> at
> org.apache.cassandra.db.transform.UnfilteredRows.add(UnfilteredRows.java:41)
> at
> org.apache.cassandra.db.transform.Transformation.add(Transformation.java:156)
> at
> org.apache.cassandra.db.transform.Transformation.apply(Transformation.java:122)
> at
> org.apache.cassandra.io.sstable.format.big.BigTableWriter.append(BigTableWriter.java:147)
> at
> org.apache.cassandra.io.sstable.SSTableRewriter.append(SSTableRewriter.java:125)
> at
> org.apache.cassandra.db.compaction.writers.DefaultCompactionWriter.realAppend(DefaultCompactionWriter.java:57)
> at
> org.apache.cassandra.db.compaction.writers.CompactionAwareWriter.append(CompactionAwareWriter.java:109)
> at
> org.apache.cassandra.db.compaction.CompactionTask.runMayThrow(CompactionTask.java:195)
> at
> org.apache.cassandra.utils.WrappedRunnable.run(WrappedRunnable.java:28)
> at
> org.apache.cassandra.db.compaction.CompactionTask.executeInternal(CompactionTask.java:89)
> at
> org.apache.cassandra.db.compaction.AbstractCompactionTask.execute(AbstractCompactionTask.java:61)
> at
> org.apache.cassandra.db.compaction.CompactionManager$5.execute(CompactionManager.java:416)
> at
> org.apache.cassandra.db.compaction.CompactionManager$2.call(CompactionManager.java:308)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at
> org.apache.cassandra.concurrent.NamedThreadFactory.lambda$0(NamedThreadFactory.java:79)
> at java.lang.Thread.run(Thread.java:745)
> {code}
> This problem is cause
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
