[ 
https://issues.apache.org/jira/browse/TRAFODION-188?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14696264#comment-14696264
 ] 

Atanu Mishra commented on TRAFODION-188:
----------------------------------------

Weishiun Tsai (wei-shiun-tsai) wrote on 2014-05-19:     #1
Verified on the datalake v40646 build. This problem has been fixed:

>>create table t0 (a int not null not droppable, b int, c int, primary key (a)) 
>>salt using 12 partitions;

--- SQL operation complete.
>>
>>create table t1 (a int not null not droppable, b int, c int, primary key (a)) 
>>salt using 12 partitions;

--- SQL operation complete.
>>
>>create table t2 (a int not null not droppable, b int, c int, primary key (a)) 
>>salt using 12 partitions;

--- SQL operation complete.
>>
>>create table t6 (a int not null not droppable, b int, c int, primary key (a)) 
>>salt using 12 partitions;

--- SQL operation complete.
>>
>>create table t8 (a int not null not droppable, b int, c int, primary key (a)) 
>>salt using 12 partitions;

--- SQL operation complete.
>>
>>create table cube1
+>(a int not null not droppable,
+>b int not null not droppable,
+>c int not null not droppable,
+>d int, e int, f int, txt char(100),
+>primary key (a,b,c))
+>store by primary key salt using 12 partitions;

--- SQL operation complete.
>>
>>insert into t0 values 
>>(0,0,0),(1,1,1),(2,2,2),(3,3,3),(4,4,4),(5,5,5),(6,6,6),(7,7,7),(8,8,8),(9,9,9);

--- 10 row(s) inserted.
>>
>>upsert using load into t1 select * from t0;

--- 10 row(s) inserted.
>>
>>upsert using load into t2 select * from t0;

--- 10 row(s) inserted.
>>
>>upsert using load into t6 select t1.a+10*t2.a,t1.a,t2.a from t1,t2;

--- 100 row(s) inserted.
>>
>>upsert using load into t8 select t6.a+100*t1.a,t6.a,t1.a from t1,t6;

--- 1000 row(s) inserted.
>>
>>insert into cube1 select t1.a, t6.a, t8.a, t1.a, t6.a, t8.a, 'some text' from 
>>t1, t6, t8 where t8.a < 100;

--- 100000 row(s) inserted.

Changed in trafodion:
status: New → Fix Released
Weishiun Tsai (wei-shiun-tsai) on 2014-05-19
summary:        - Insert..select returns HBASE_ACCSES_ERROR(-705) and then TMF 
error 97 in
+ Insert..select returns HBASE_ACCESS_ERROR(-705) and then TMF error 97 in
org.apache.hadoop.hbase.regionserver.transactional

> LP Bug: 1286349 - Insert..select returns HBASE_ACCESS_ERROR(-705) and then 
> TMF error 97 in org.apache.hadoop.hbase.regionserver.transactional
> ---------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: TRAFODION-188
>                 URL: https://issues.apache.org/jira/browse/TRAFODION-188
>             Project: Apache Trafodion
>          Issue Type: Bug
>          Components: dtm
>            Reporter: Weishiun Tsai
>            Priority: Critical
>
> A insert..select statement returns HBASE_ACCSES_ERROR(-705) in 
> org.apache.hadoop.hbase.regionserver.transactional and then TMF error 97 
> afterwards, as shown here.  This problem is fairly reproducible now with the 
> QA test to create and populate hcubedb through sqlci/JDBC.  This may or may 
> not be related to https://bugs.launchpad.net/trafodion/+bug/1274716 “Getting 
> TM error 97 for INSERT SELECT <hive table>”.  In this case, all tables from 
> the select list are Trafodion/hbase tables, but in the other case, the select 
> table is a hive table.  For now, we use this case to track the problem 
> separately.
> SQL>insert into cube1 select t1.a, t6.a, t8.a, t1.a, t6.a, t8.a, 'some text'
> from t1, t6, t8 where t8.a < 100;
> *** ERROR[8448] Unable to access Hbase interface. Call to 
> ExpHbaseInterface::rowExists returned error HBASE_ACCESS_ERROR(-705). Cause:
> org.apache.hadoop.hbase.client.transactional.UnknownTransactionException: 
> org.apache.hadoop.hbase.client.transactional.UnknownTransactionException: 
> transaction: [4294996353], region: 
> [TRAFODION.G_HCUBEDB.CUBE1,,1393442223602.979ac8a950b9df29d1ab4832dd9acb37.]
>         at 
> org.apache.hadoop.hbase.regionserver.transactional.TransactionalRegion.getTransactionState(TransactionalRegion.java:741)
>         at 
> org.apache.hadoop.hbase.regionserver.transactional.TransactionalRegion.checkAndPut(TransactionalRegion.java:941)
>         at 
> org.apache.hadoop.hbase.regionserver.transactional.TransactionalRegionServer.checkAndPut(TransactionalRegionServer.java:449)
>         at sun.reflect.GeneratedMethodAccessor39.invoke(Unknown Source)
>         at 
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
>         at java.lang.reflect.Method.invoke(Method.java:597)
>         at 
> org.apache.hadoop.hbase.ipc.WritableRpcEngine$Server.call(WritableRpcEngine.java:320)
>         at 
> org.apache.hadoop.hbase.ipc.HBaseServer$Handler.run(HBaseServer.java:1428)
> sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
> sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:57)
> sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
> java.lang.reflect.Constructor.newInstance(Constructor.java:526)
> org.apache.hadoop.ipc.RemoteException.instantiateException(RemoteException.java:90)
> org.apache.hadoop.ipc.RemoteException.unwrapRemoteException(RemoteException.java:79)
> org.apache.hadoop.hbase.client.ServerCallable.translateException(ServerCallable.java:228)
> org.apache.hadoop.hbase.client.ServerCallable.withRetries(ServerCallable.java:166)
> org.apache.hadoop.hbase.client.transactional.TransactionalTable.checkAndPut(TransactionalTable.java:212)
> org.apache.hadoop.hbase.client.transactional.RMInterface.checkAndPut(RMInterface.java:164)
> org.trafodion.sql.HBaseAccess.HTableClient.checkAndInsertRow(HTableClient.java:537)
> . [2014-02-26 11:36:31]
> *** ERROR[8610] Transaction subsystem TMF reported error 97 on a waited 
> rollback transaction. [2014-02-26 11:36:31]
> This problem has been reproduced on rhel-cdh1 with the following TM tracing 
> and hbase tracing turned on.  Narendra has collected the tracing info for 
> further debugging.
> 1) Change the logging threshold of the HBase Region Server from ‘INFO’ to 
> ‘DEBUG’: Via the Cloudera Manager:
>                                                      i. Click on HBase 
> Service->Configuration tab -> ‘View and Edit’
> 1. Click on RegionServer -> Logs -> Logging Threshold to ‘DEBUG’
>                                                    ii. Restart HBase
> 2) Trace the Trafodion TM:
>  a. In sqf/etc/ms.env (then pdcp to all the nodes):
> # DTM Tracing. To enable standard tracing uncomment the following lines.
>  TM_TRACE=1
>  HBASE_TRACE=0xf
>  TM_TRACE_DETAIL=3
>  TM_TRACE_UNIQUE=1
>  TM_TRACE_FILE=zzttm.
>  b. Restart SQ
> 3) In the file $MY_SQROOT/logs/log4j.dtm.config (all nodes, log file: 
> $MY_SQROOT/sql/scripts/trafodion.dtm.log):
> log4j.logger.org.trafodion.dtm=TRACE (change it from DEBUG to TRACE)



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to