Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cassandra.apache.org
Date: Tue, 29 Aug 2017 12:10:03 +0000 (UTC)
From: "Marcus Eriksson (JIRA)" <jira@apache.org>
To: commits@cassandra.apache.org
Message-ID: <JIRA.13093731.1502359406000.145719.1504008603232@Atlassian.JIRA>
In-Reply-To: <JIRA.13093731.1502359406000@Atlassian.JIRA>
References: <JIRA.13093731.1502359406000@Atlassian.JIRA> <JIRA.13093731.1502359406616@jira-lw-us.apache.org>
Subject: [jira] [Commented] (CASSANDRA-13752) Corrupted SSTables created in
 3.11
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Tue, 29 Aug 2017 12:10:12 -0000


    [ https://issues.apache.org/jira/browse/CASSANDRA-13752?page=3Dcom.atla=
ssian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=
=3D16145183#comment-16145183 ]=20

Marcus Eriksson commented on CASSANDRA-13752:
---------------------------------------------

So this is because we reuse the same StreamingHistogram when we [open sstab=
les early|https://github.com/apache/cassandra/blob/cassandra-3.11/src/java/=
org/apache/cassandra/io/sstable/metadata/MetadataCollector.java#L292] - the=
 early opened sstable will be using the streaming histogram that compaction=
 is still building. Since CASSANDRA-13038 we can modify the contents of the=
 StreamingHistogram when we call {{sum()}} ({{spool}} might be compacted in=
to {{bin}}). So, if someone calls the {{ColumnFamilyStoreMBean#getDroppable=
TombstoneRatio}} at the wrong time we could get either the CME from CASSAND=
RA-13756 or this corruption.

Making StreamingHistogram thread safe is one way of fixing this, but I woul=
d argue that we should be using a "builder" for StreamingHistogram - we sho=
uld never access the SH while building it and for early opened sstables we =
should call {{.build()}} on the StreamingHistogramBuilder and get a copy of=
 the internal state.

Also, we should not query LIVE sstables [here|https://github.com/apache/cas=
sandra/blob/cassandra-3.11/src/java/org/apache/cassandra/db/ColumnFamilySto=
re.java#L2589] - it should be using {{SSTableSet.CANONICAL}} (this is proba=
bly enough to fix this for now - this is the only way I can see that we acc=
ess the sstablemetadata in early opened sstables).

> Corrupted SSTables created in 3.11
> ----------------------------------
>
>                 Key: CASSANDRA-13752
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1375=
2
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Hannu Kr=C3=B6ger
>            Assignee: Hannu Kr=C3=B6ger
>            Priority: Blocker
>             Fix For: 3.11.1
>
>
> We have discovered issues with corrupted SSTables.=20
> {code}
> ERROR [SSTableBatchOpen:22] 2017-08-03 20:19:53,195 SSTableReader.java:57=
7 - Cannot read sstable /cassandra/data/mykeyspace/mytable-7a4992800d5611e7=
b782cb90016f2d17/mc-35556-big=3D[Data.db, Statistics.db, Summary.db, Digest=
.crc32, CompressionInfo.db, TOC.txt, Index.db, Filter.db]; other IO error, =
skipping table
> java.io.EOFException: EOF after 1898 bytes out of 21093
>         at org.apache.cassandra.io.util.RebufferingInputStream.readFully(=
RebufferingInputStream.java:68) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at org.apache.cassandra.io.util.RebufferingInputStream.readFully(=
RebufferingInputStream.java:60) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at org.apache.cassandra.utils.ByteBufferUtil.read(ByteBufferUtil.=
java:402) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at org.apache.cassandra.utils.ByteBufferUtil.readWithShortLength(=
ByteBufferUtil.java:377) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at org.apache.cassandra.io.sstable.metadata.StatsMetadata$StatsMe=
tadataSerializer.deserialize(StatsMetadata.java:325) ~[apache-cassandra-3.1=
1.0.jar:3.11.0]
>         at org.apache.cassandra.io.sstable.metadata.StatsMetadata$StatsMe=
tadataSerializer.deserialize(StatsMetadata.java:231) ~[apache-cassandra-3.1=
1.0.jar:3.11.0]
>         at org.apache.cassandra.io.sstable.metadata.MetadataSerializer.de=
serialize(MetadataSerializer.java:122) ~[apache-cassandra-3.11.0.jar:3.11.0=
]
>         at org.apache.cassandra.io.sstable.metadata.MetadataSerializer.de=
serialize(MetadataSerializer.java:93) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at org.apache.cassandra.io.sstable.format.SSTableReader.open(SSTa=
bleReader.java:488) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at org.apache.cassandra.io.sstable.format.SSTableReader.open(SSTa=
bleReader.java:396) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at org.apache.cassandra.io.sstable.format.SSTableReader$5.run(SST=
ableReader.java:561) ~[apache-cassandra-3.11.0.jar:3.11.0]
>         at java.util.concurrent.Executors$RunnableAdapter.call(Executors.=
java:511) [na:1.8.0_111]
>         at java.util.concurrent.FutureTask.run(FutureTask.java:266) [na:1=
.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolEx=
ecutor.java:1142) [na:1.8.0_111]
>         at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolE=
xecutor.java:617) [na:1.8.0_111]
>         at org.apache.cassandra.concurrent.NamedThreadFactory.lambda$thre=
adLocalDeallocator$0(NamedThreadFactory.java:81) [apache-cassandra-3.11.0.j=
ar:3.11.0]
> {code}
> Files look like this:
> {code}
> -rw-r--r--. 1 cassandra cassandra     3899251 Aug  7 08:37 mc-6166-big-Co=
mpressionInfo.db
> -rw-r--r--. 1 cassandra cassandra 16874421686 Aug  7 08:37 mc-6166-big-Da=
ta.db
> -rw-r--r--. 1 cassandra cassandra          10 Aug  7 08:37 mc-6166-big-Di=
gest.crc32
> -rw-r--r--. 1 cassandra cassandra     2930904 Aug  7 08:37 mc-6166-big-Fi=
lter.db
> -rw-r--r--. 1 cassandra cassandra   111175880 Aug  7 08:37 mc-6166-big-In=
dex.db
> -rw-r--r--. 1 cassandra cassandra       13762 Aug  7 08:37 mc-6166-big-St=
atistics.db
> -rw-r--r--. 1 cassandra cassandra      882008 Aug  7 08:37 mc-6166-big-Su=
mmary.db
> -rw-r--r--. 1 cassandra cassandra          92 Aug  7 08:37 mc-6166-big-TO=
C.txt
> {code}


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org