You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Billy Pearson (JIRA)" <ji...@apache.org> on 2009/08/18 21:10:14 UTC
[jira] Created: (HBASE-1777) column length is not checked before
saved to memstore
column length is not checked before saved to memstore
-----------------------------------------------------
Key: HBASE-1777
URL: https://issues.apache.org/jira/browse/HBASE-1777
Project: Hadoop HBase
Issue Type: Bug
Components: client
Affects Versions: 0.20.0
Reporter: Billy Pearson
Fix For: 0.20.1
I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
we should check this before saving the record to memstore
As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
{code}
2009-08-18 12:54:16,572 FATAL
org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
required. Forcing server shutdown
org.apache.hadoop.hbase.DroppedSnapshotException: region:
webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
Caused by: java.io.IOException: Key length 183108 > 65536
at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
... 3 more
{code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "Amandeep Khurana (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amandeep Khurana updated HBASE-1777:
------------------------------------
Assignee: Amandeep Khurana
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.20.1
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "Amandeep Khurana (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amandeep Khurana updated HBASE-1777:
------------------------------------
Assignee: Andrew Purtell (was: Amandeep Khurana)
Status: Patch Available (was: Open)
Attaching a patch. Havent got a chance to write a test case for this yet.
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Andrew Purtell
> Fix For: 0.21.0
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
stack updated HBASE-1777:
-------------------------
Moving out of 0.20.1 after chatting with Aman. Its rare that we run into this issue -- though its bad when it happens.
A while back, looking at this issue with Aman, it looks like there are some interesting mismatches in how length checking is done as you go down from client through server and out to hfile so its needs a bit of careful study.
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.21.0
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "Jonathan Gray (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Gray updated HBASE-1777:
---------------------------------
Resolution: Fixed
Status: Resolved (was: Patch Available)
Committed to branch and trunk.
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.20.2, 0.21.0
>
> Attachments: HBASE-1777.patch
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "Amandeep Khurana (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amandeep Khurana updated HBASE-1777:
------------------------------------
Fix Version/s: (was: 0.20.1)
0.21.0
Moving it to 0.21 since I didnt get time to look into it yet.
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.21.0
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "Amandeep Khurana (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amandeep Khurana updated HBASE-1777:
------------------------------------
Attachment: HBASE-1777.patch
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Andrew Purtell
> Fix For: 0.21.0
>
> Attachments: HBASE-1777.patch
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "stack (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12744659#action_12744659 ]
stack commented on HBASE-1777:
------------------------------
Good one Billy. Lets fix in 0.20.1.
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Fix For: 0.20.1
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "Jonathan Gray (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Gray updated HBASE-1777:
---------------------------------
Fix Version/s: 0.20.2
Moving back into 0.20.2... This got moved out and never committed
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.20.2, 0.21.0
>
> Attachments: HBASE-1777.patch
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "ryan rawson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12763697#action_12763697 ]
ryan rawson commented on HBASE-1777:
------------------------------------
this one is not as straightforward as it seems, the mainline code in Put and KeyValue checks the row length, thus this is "impossible". Must be one of the other code paths somewhere...
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.21.0
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "Amandeep Khurana (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amandeep Khurana reassigned HBASE-1777:
---------------------------------------
Assignee: Amandeep Khurana (was: Andrew Purtell)
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.21.0
>
> Attachments: HBASE-1777.patch
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HBASE-1777) column length is not checked before
saved to memstore
Posted by "ryan rawson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HBASE-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12763709#action_12763709 ]
ryan rawson commented on HBASE-1777:
------------------------------------
ah looks like its a minor bug in HFile, the code is:
if (length > MAXIMUM_KEY_LENGTH) {
throw new IOException("Key length " + length + " > " +
MAXIMUM_KEY_LENGTH);
}
MAXIMUM_KEY_LENGTH is 64*1024, it should be really (2^31-1), since a HFile key can be up to that sized. The actual row portion can only be 32/64k.
> column length is not checked before saved to memstore
> -----------------------------------------------------
>
> Key: HBASE-1777
> URL: https://issues.apache.org/jira/browse/HBASE-1777
> Project: Hadoop HBase
> Issue Type: Bug
> Components: client
> Affects Versions: 0.20.0
> Reporter: Billy Pearson
> Assignee: Amandeep Khurana
> Fix For: 0.21.0
>
>
> I added some debuging to line 511 in HFile.java and found that the column is causing my problem it was > max size
> we should check this before saving the record to memstore
> As of 0.20.0-RC2 the server dies and cause the hlogs to be read again by the next region server that gets the region in the end it cause the whole cluster to go down sense the bad data is in the hlog at this point.
> {code}
> 2009-08-18 12:54:16,572 FATAL
> org.apache.hadoop.hbase.regionserver.MemStoreFlusher: Replay of hlog
> required. Forcing server shutdown
> org.apache.hadoop.hbase.DroppedSnapshotException: region:
> webdata,http:\x2F\x2Fanaal-genomen.isporno.nl\x2F,1250569930062
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:950)
> at org.apache.hadoop.hbase.regionserver.HRegion.flushcache(HRegion.java:843)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.flushRegion(MemStoreFlusher.java:241)
> at org.apache.hadoop.hbase.regionserver.MemStoreFlusher.run(MemStoreFlusher.java:149)
> Caused by: java.io.IOException: Key length 183108 > 65536
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.checkKey(HFile.java:511)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:479)
> at org.apache.hadoop.hbase.io.hfile.HFile$Writer.append(HFile.java:447)
> at org.apache.hadoop.hbase.regionserver.Store.internalFlushCache(Store.java:525)
> at org.apache.hadoop.hbase.regionserver.Store.flushCache(Store.java:489)
> at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:935)
> ... 3 more
> {code}
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.