You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Shashank Thillainathan (Jira)" <ji...@apache.org> on 2021/04/29 10:57:00 UTC
[jira] [Updated] (HBASE-25827) Per Cell TTL tags get duplicated
with increments causing tags length overflow
[ https://issues.apache.org/jira/browse/HBASE-25827?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Shashank Thillainathan updated HBASE-25827:
-------------------------------------------
Description:
Incrementing with per cell TTL and flushing corrupts the HFile.
Reproducing the issue:
Incrementing a row and a column with per cell TTL for about 3 thousand times and flushing corrupts the HFile leaving the table unusable.
Cause:
On reading the HFile, it is found that duplicate TTL tags get appended for each cell.
Though this case has already been addressed here at HBASE-18030, corruption still occurs even with this patch.
{code:java}
java.lang.IllegalStateException: Invalid currTagsLen -31260. Block offset: 250962, block length: 76568, position: 42207 (without header). path=hdfs://hdfs/file/path
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.checkTagsLen(HFileReaderImpl.java:642)
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.readKeyValueLen(HFileReaderImpl.java:630)
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl._next(HFileReaderImpl.java:1080)
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.next(HFileReaderImpl.java:1097)
at org.apache.hadoop.hbase.regionserver.StoreFileScanner.next(StoreFileScanner.java:208)
at org.apache.hadoop.hbase.regionserver.KeyValueHeap.next(KeyValueHeap.java:120)
at org.apache.hadoop.hbase.regionserver.StoreScanner.next(StoreScanner.java:654)
at org.apache.hadoop.hbase.regionserver.compactions.Compactor.performCompaction(Compactor.java:388)
at org.apache.hadoop.hbase.regionserver.compactions.Compactor.compact(Compactor.java:327)
at org.apache.hadoop.hbase.regionserver.compactions.DefaultCompactor.compact(DefaultCompactor.java:65)
at org.apache.hadoop.hbase.regionserver.DefaultStoreEngine$DefaultCompactionContext.compact(DefaultStoreEngine.java:126)
at org.apache.hadoop.hbase.regionserver.HStore.compact(HStore.java:1432)
at org.apache.hadoop.hbase.regionserver.HRegion.compact(HRegion.java:2192)
at org.apache.hadoop.hbase.regionserver.CompactSplit$CompactionRunner.doCompaction(CompactSplit.java:577)
at org.apache.hadoop.hbase.regionserver.CompactSplit$CompactionRunner.run(CompactSplit.java:619)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
{code}
was:
Incrementing with per cell TTL and flushing corrupts the HFile.
Reproducing the issue:
Incrementing a row and a column with per cell TTL for about 3 thousand times and flushing corrupts the HFile leaving the table unusable.
Cause:
On reading the HFile, it is found that duplicate TTL tags get appended for each cell.
Though this case has already been addressed here at [HBASE-18030|https://issues.apache.org/jira/browse/HBASE-18030], corruption still occurs even with this patch.
{code:java}
// java.lang.IllegalStateException: Invalid currTagsLen -31260. Block offset: 250962, block length: 76568, position: 42207 (without header). path=hdfs://hdfs/file/path
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.checkTagsLen(HFileReaderImpl.java:642)
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.readKeyValueLen(HFileReaderImpl.java:630)
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl._next(HFileReaderImpl.java:1080)
at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.next(HFileReaderImpl.java:1097)
at org.apache.hadoop.hbase.regionserver.StoreFileScanner.next(StoreFileScanner.java:208)
at org.apache.hadoop.hbase.regionserver.KeyValueHeap.next(KeyValueHeap.java:120)
at org.apache.hadoop.hbase.regionserver.StoreScanner.next(StoreScanner.java:654)
at org.apache.hadoop.hbase.regionserver.compactions.Compactor.performCompaction(Compactor.java:388)
at org.apache.hadoop.hbase.regionserver.compactions.Compactor.compact(Compactor.java:327)
at org.apache.hadoop.hbase.regionserver.compactions.DefaultCompactor.compact(DefaultCompactor.java:65)
at org.apache.hadoop.hbase.regionserver.DefaultStoreEngine$DefaultCompactionContext.compact(DefaultStoreEngine.java:126)
at org.apache.hadoop.hbase.regionserver.HStore.compact(HStore.java:1432)
at org.apache.hadoop.hbase.regionserver.HRegion.compact(HRegion.java:2192)
at org.apache.hadoop.hbase.regionserver.CompactSplit$CompactionRunner.doCompaction(CompactSplit.java:577)
at org.apache.hadoop.hbase.regionserver.CompactSplit$CompactionRunner.run(CompactSplit.java:619)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
{code}
> Per Cell TTL tags get duplicated with increments causing tags length overflow
> -----------------------------------------------------------------------------
>
> Key: HBASE-25827
> URL: https://issues.apache.org/jira/browse/HBASE-25827
> Project: HBase
> Issue Type: Bug
> Components: regionserver
> Affects Versions: 2.1.9, 2.2.6
> Reporter: Shashank Thillainathan
> Priority: Critical
>
> Incrementing with per cell TTL and flushing corrupts the HFile.
> Reproducing the issue:
> Incrementing a row and a column with per cell TTL for about 3 thousand times and flushing corrupts the HFile leaving the table unusable.
>
> Cause:
> On reading the HFile, it is found that duplicate TTL tags get appended for each cell.
> Though this case has already been addressed here at HBASE-18030, corruption still occurs even with this patch.
> {code:java}
> java.lang.IllegalStateException: Invalid currTagsLen -31260. Block offset: 250962, block length: 76568, position: 42207 (without header). path=hdfs://hdfs/file/path
> at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.checkTagsLen(HFileReaderImpl.java:642)
> at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.readKeyValueLen(HFileReaderImpl.java:630)
> at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl._next(HFileReaderImpl.java:1080)
> at org.apache.hadoop.hbase.io.hfile.HFileReaderImpl$HFileScannerImpl.next(HFileReaderImpl.java:1097)
> at org.apache.hadoop.hbase.regionserver.StoreFileScanner.next(StoreFileScanner.java:208)
> at org.apache.hadoop.hbase.regionserver.KeyValueHeap.next(KeyValueHeap.java:120)
> at org.apache.hadoop.hbase.regionserver.StoreScanner.next(StoreScanner.java:654)
> at org.apache.hadoop.hbase.regionserver.compactions.Compactor.performCompaction(Compactor.java:388)
> at org.apache.hadoop.hbase.regionserver.compactions.Compactor.compact(Compactor.java:327)
> at org.apache.hadoop.hbase.regionserver.compactions.DefaultCompactor.compact(DefaultCompactor.java:65)
> at org.apache.hadoop.hbase.regionserver.DefaultStoreEngine$DefaultCompactionContext.compact(DefaultStoreEngine.java:126)
> at org.apache.hadoop.hbase.regionserver.HStore.compact(HStore.java:1432)
> at org.apache.hadoop.hbase.regionserver.HRegion.compact(HRegion.java:2192)
> at org.apache.hadoop.hbase.regionserver.CompactSplit$CompactionRunner.doCompaction(CompactSplit.java:577)
> at org.apache.hadoop.hbase.regionserver.CompactSplit$CompactionRunner.run(CompactSplit.java:619)
> at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> {code}
>
>
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)