[
https://issues.apache.org/jira/browse/HBASE-1500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12718183#action_12718183
]
Lars George commented on HBASE-1500:
------------------------------------
It looks like it happens here from the above
{code}
public int compare(byte[] left, int loffset, int llength, byte[] right,
int roffset, int rlength) {
// Compare row
short lrowlength = Bytes.toShort(left, loffset);
short rrowlength = Bytes.toShort(right, roffset);
int compare = compareRows(left, loffset + Bytes.SIZEOF_SHORT,
lrowlength,
right, roffset + Bytes.SIZEOF_SHORT, rrowlength);
if (compare != 0) {
return compare;
}
{code}
It reads the lengths from the key byte array assuming they are correct and then
applies those lengths to the compareRow() call, which then fails.
This indicates a borked length from a key array? Could you instrument the code
above and log the lrowlength, rrowlength and the the real left.length and
right.length?
If that mismatches, the question is why?
> KeyValue$KeyComparator array overrun
> ------------------------------------
>
> Key: HBASE-1500
> URL: https://issues.apache.org/jira/browse/HBASE-1500
> Project: Hadoop HBase
> Issue Type: Bug
> Affects Versions: 0.20.0
> Reporter: Andrew Purtell
> Priority: Blocker
> Fix For: 0.20.0
>
>
> {code}
> 09/06/08 22:58:47 INFO zookeeper.ZooKeeper: Initiating client connection,
> host=B
> OA03:2181,BOA02:2181,BOA01:2181,BOA04:2181 sessionTimeout=10000
> watcher=org.apac
> he.hadoop.hbase.zookeeper.watcherwrap...@518bf072
> 09/06/08 22:58:47 INFO zookeeper.ClientCnxn: zookeeper.disableAutoWatchReset
> is
> false
> 09/06/08 22:58:47 INFO zookeeper.ClientCnxn: Attempting connection to server
> BOA
> 04/172.20.3.231:2181
> 09/06/08 22:58:47 INFO zookeeper.ClientCnxn: Priming connection to
> java.nio.chan
> nels.SocketChannel[connected local=/172.20.3.232:40296
> remote=BOA04/172.20.3.231
> :2181]
> 09/06/08 22:58:47 INFO zookeeper.ClientCnxn: Server connection successful
> 09/06/08 22:58:47 WARN mapred.JobClient: Use GenericOptionsParser for parsing
> th
> e arguments. Applications should implement Tool for the same.
> 09/06/08 22:58:47 WARN mapred.JobClient: No job jar file set. User classes
> may
> not be found. See JobConf(Class) or JobConf#setJar(String).
> 09/06/08 22:58:47 INFO zookeeper.ZooKeeper: Initiating client connection,
> host=B
> OA03:2181,BOA02:2181,BOA01:2181,BOA04:2181 sessionTimeout=10000
> watcher=org.apac
> he.hadoop.hbase.zookeeper.watcherwrap...@362f0d54
> 09/06/08 22:58:47 INFO zookeeper.ClientCnxn: Attempting connection to server
> BOA
> 03/172.20.3.230:2181
> 09/06/08 22:58:47 INFO zookeeper.ClientCnxn: Priming connection to
> java.nio.chan
> nels.SocketChannel[connected local=/172.20.3.232:42792
> remote=BOA03/172.20.3.230
> :2181]
> 09/06/08 22:58:47 INFO zookeeper.ClientCnxn: Server connection successful
> 09/06/08 22:58:48 INFO mapred.TableInputFormatBase: split:
> 0->BOA04.trendmicro.c
> om:,01e33c601a7a9dd0ddb5c8427438f2f1
> Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 32
> at org.apache.hadoop.hbase.util.Bytes.compareTo(Bytes.java:798)
> at
> org.apache.hadoop.hbase.KeyValue$KeyComparator.compareRows(KeyValue.j
> ava:1760)
> at
> org.apache.hadoop.hbase.KeyValue$KeyComparator.compare(KeyValue.java:
> 1696)
> at
> org.apache.hadoop.hbase.KeyValue$KeyComparator.compare(KeyValue.java:
> 1755)
> at
> org.apache.hadoop.hbase.KeyValue$KeyComparator.compare(KeyValue.java:
> 1687)
> at
> org.apache.hadoop.hbase.client.HConnectionManager$TableServers.getCac
> hedLocation(HConnectionManager.java:697)
> at
> org.apache.hadoop.hbase.client.HConnectionManager$TableServers.locate
> RegionInMeta(HConnectionManager.java:541)
> at
> org.apache.hadoop.hbase.client.HConnectionManager$TableServers.locate
> Region(HConnectionManager.java:525)
> at
> org.apache.hadoop.hbase.client.HConnectionManager$TableServers.locate
> Region(HConnectionManager.java:488)
> at
> org.apache.hadoop.hbase.client.HConnectionManager$TableServers.getReg
> ionLocation(HConnectionManager.java:342)
> at
> org.apache.hadoop.hbase.client.HTable.getRegionLocation(HTable.java:1
> 91)
> at
> org.apache.hadoop.hbase.mapred.TableInputFormatBase.getSplits(TableIn
> putFormatBase.java:296)
> at org.apache.hadoop.mapred.JobClient.submitJob(JobClient.java:742)
> at org.apache.hadoop.mapred.JobClient.runJob(JobClient.java:1026)
> at net.iridiant.crawler.mapred.DocumentParser.main(Unknown Source)
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
