[
https://issues.apache.org/jira/browse/CASSANDRA-20190?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17938608#comment-17938608
]
Dmitry Konstantinov commented on CASSANDRA-20190:
-------------------------------------------------
A possible way can be to add a heuristic check into
org.apache.cassandra.io.sstable.indexsummary.IndexSummary.IndexSummarySerializer#deserialize
once we loaded offsets and entries into memory we can read offsets.getInt(0)
(and/or entries.getLong(..)) and check if it is suspiciously big for example by
offsets.getInt(0) > Integer.reserveBytes(offsets.getInt(0) or a similar way and
then we can throw IOException.
We load summaries using the following call tree if I read the code correctly:
*
org.apache.cassandra.io.sstable.format.big.BigSSTableReaderLoadingBuilder#openComponents
<-- it should rebuild the summary
*
org.apache.cassandra.io.sstable.format.big.BigSSTableReaderLoadingBuilder#loadSummary
<-- it will delete the summary file
*
org.apache.cassandra.io.sstable.format.big.IndexSummaryComponent#loadOrDeleteCorrupted
<-- it will catch the IOException
* org.apache.cassandra.io.sstable.format.big.IndexSummaryComponent#load
*
org.apache.cassandra.io.sstable.indexsummary.IndexSummary.IndexSummarySerializer#deserialize
so, it looks like it would be the cheapest way to do such detection...
> MemoryUtil.setInt/getInt and similar use the wrong endianness
> -------------------------------------------------------------
>
> Key: CASSANDRA-20190
> URL: https://issues.apache.org/jira/browse/CASSANDRA-20190
> Project: Apache Cassandra
> Issue Type: Bug
> Components: Local/Other
> Reporter: Branimir Lambov
> Assignee: Dmitry Konstantinov
> Priority: Normal
> Fix For: 5.x
>
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> `NativeCell`, `NativeClustering` and `NativeDecoratedKey` use the above
> methods from `MemoryUtil` to write and read data from native memory. As far
> as I can see they are meant to write data in big endian. They do not (they
> always correct to little endian).
> Moreover, they disagree with their `ByByte` versions on big-endian machines
> (which is only likely an issue on aligned-access architectures (x86 and arm
> should be fine)).
> The same is true for the methods in `Memory`, used by compression metadata as
> well as index summaries.
> We need to verify that this does not cause any problems, and to change the
> methods to behave as expected and document the behaviour by explicitly using
> `ByteOrder.LITTLE_ENDIAN` for any data that may have been persisted on disk
> with the wrong endianness.
>
> The current MemoryUtil behaviour (before the fix):
> ||Native
> order||MemoryUtil.setX||MemoryUtil.setXByByte||MemoryUtil.getX||MemoryUtil.getXByByte||
> |BE|LE|BE|LE|BE|
> |LE|LE|LE|LE|LE|
> shortly: MemoryUtil.setX/getX is LE, MemoryUtil.setXByByte/getXByByte is
> Native
--
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
