Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Sun, 23 Mar 2014 11:40:43 +0000 (UTC)
From: "LiJianyin (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.12702613.1395315439391.126864.1395574843686@arcas>
In-Reply-To: <JIRA.12702613.1395315439391@arcas>
References: <JIRA.12702613.1395315439391@arcas>
Subject: [jira] [Commented] (HBASE-10803) "mutateRow" maybe fail when Region
 Server restarted(or killed) or Region Balance
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/HBASE-10803?page=3Dcom.atlassia=
n.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D139=
44407#comment-13944407 ]=20

LiJianyin commented on HBASE-10803:
-----------------------------------

Thanks, Lars Hofhansl, stack. I'll try to write  a unit tests.
in addition=EF=BC=9A

=09two column family: CF1, CF0
 =09addOrUpdate: put->CF1 with the timestamp
 =09delete(mutateRow): delete->CF1(deleteColumns) and put->CF0 with the sam=
e timestamp
 =09all timestamps are from client (System.currentTimeMillis())
=09
=09so, if delete(mutateRow) success, the CF0 must have value.  but the resu=
lt is not so.
=09
 =09for example: (rowkey+qualifier only in DEL_SUCCESS_LOG)
 =09=09in the DEL_SUCCESS_LOG:
=09=09=09 timestamp:1394601911543  rowkey:361523835735633 qualifier:3670583=
952628921 value:0
=09=09--result: there is no value in CF0 with the rowkey+qualifier
=09=09
         =09another example: (the same rowkey+qualifier are both in DEL_SUC=
CESS_LOG and PUT_SUCCESS_LOG)
=09=09in the PUT_SUCCESS_LOG:
=09=09=09timestamp:1394601634849  rowkey:213772920506212 qualifier:36872888=
54440310 value:0
=09=09in the DEL_SUCCESS_LOG:
=09=09=09timestamp:1394601913327  rowkey:213772920506212 qualifier:36872888=
54440310 value:0
=09=09--result: CF1 with the rowkey+qualifier+timestamp 1394601634849 has t=
he value, but no value in CF0  with the rowkey+qualifier

> "mutateRow" maybe fail when Region Server restarted(or killed) or Region =
Balance
> -------------------------------------------------------------------------=
-------
>
>                 Key: HBASE-10803
>                 URL: https://issues.apache.org/jira/browse/HBASE-10803
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 0.96.1.1
>         Environment: hbase0.96.1.1
> hadoop2.2=20
> jdk1.7
> 5 Region Server=20
>            Reporter: LiJianyin
>
>  found that the operation "mutateRow" maybe fail if Region Server restart=
ed(or killed) or Region Balance, but the java client didnot throw exception=
s and  there was no exception in the server's log either.=20
> the client code :
> =09=09=09HTableInterface table =3D connection.getTable(TABLE_NAME);
> =09=09=09Delete d =3D new Delete(Bytes.toBytes(rowkey), time);
> =09=09=09d.deleteColumns(FAMILY, Bytes.toBytes(id), time);
> =09=09=09Put p =3D new Put(Bytes.toBytes(rowkey));
> =09=09=09p.add(FAMILY_DEL, Bytes.toBytes(id), time, Bytes.toBytes(value))=
;
> =09=09=09RowMutations rm =3D new RowMutations(Bytes.toBytes(rowkey));
> =09=09=09rm.add(d);
> =09=09=09rm.add(p);
> =09=09=09table.mutateRow(rm);
> =09=09=09table.close();
> the test data:
> =09Operator Type: put; mutateRow(delete+put)=20
> =09Operator Total Count:202194822  put:160866148 mutateRow:41328674
> =09All put correct, but 161 mutateRow incorrect (161 data was not mutateR=
owed in fact).
> =09then we checked log when the time mutateRow failed, NO EXCEPTION THROW=
ED, found that as follow:
> =09
> 39 failed  hbase auto region-balance
> 44 failed  hbase region server killed
> 59 failed  hbase region server restart
> 19 failed  hadoop datanode restart


--
This message was sent by Atlassian JIRA
(v6.2#6252)