[Mondrian] BitKey pos exceeds capacity

Hilario Fernandes hilario.fernandes at cortex-intelligence.com
Mon Nov 18 05:45:57 EST 2013


I've made the bug report: http://jira.pentaho.com/browse/MONDRIAN-1810


On Fri, Nov 15, 2013 at 7:05 PM, Hilario Fernandes <
hilario.fernandes at cortex-intelligence.com> wrote:

> Thanks for your input, i'll fill a bug report with all the information.
>
> So, you think that the schema change on the single connection used by the
> application might cause the problem. Maybe if i create a connection per
> query would help? Basically i'd be using a pool of olap connections... But
> I've seen somewhere that creating multiple olapconnections to wasn't the
> best way to go... I'll give it a try nonetheless!
>
>
> On Fri, Nov 15, 2013 at 6:01 PM, Julian Hyde <julianhyde at gmail.com> wrote:
>
>> Can you log a jira case for this? We fixed the previous issue, and your
>> scenario is sufficiently different that it's probably a separate bug.
>>
>> I wonder whether the multi-threaded nature of your application is causing
>> this. If the list of columns in a star is changing, say because one
>> connection is using a star for a query while another one is building a
>> schema and adding columns to the same star, then that might explain it. I
>> don't have the solution to the bug, but hopefully these remarks would give
>> an engineer a clue where to look.
>>
>> Julian
>>
>> On Nov 15, 2013, at 3:43 AM, Hilario Fernandes <
>> hilario.fernandes at cortex-intelligence.com> wrote:
>>
>> So, i've been trying to isolate this issue and it has not been easy, but
>> finally i've managed to get one run where i got the erros in a log i could
>> use to present here. One thing to point out, this is not consistent. The
>> sequence of queries that gave the error might not give it if i run it again
>> on the same conditions with no change in the data. This is what is making
>> it so hard to figure what's causing it.
>>
>> The project i'm working is built uppon grails framework, there is a
>> service in it that holds the olapConnection to run the queries. These are
>> the startup parameters:
>>
>> def connectionProperties = '' << 'jdbc:mondrian:'
>> connectionProperties << "Jdbc="jdbc:monetdb://host:50000/database;"
>> connectionProperties << "JdbcUser=username;"
>> connectionProperties << "JdbcPassword=password;"
>> connectionProperties << "JdbcDrivers=nl.cwi.monetdb.jdbc.MonetDriver;"
>> connectionProperties << "UseContentChecksum=true;"
>> connectionProperties << "CatalogContent=<Schema name=\"CUBES\"></Schema>"
>>
>> Class.forName("mondrian.olap4j.MondrianOlap4jDriver")
>> def connection = DriverManager.getConnection(connectionProperties)
>> def olapConnection = connection.unwrap(OlapConnection.class)
>>
>>
>> The cube schema is generated per analysis so we add it before the query
>> and remove it after its completed ("INFO  mondrian.MondrianBridge CURRENT
>> CUBES:[]" in the log indicates the cubes in the global schema at that
>> time). This is the code we use for that:
>>
>> RolapSchema schema = getServerSchema(olapConnection)
>> schema.createCube(xml)
>> ...
>> schema.removeCube(cubeId)
>>
>>
>> Query execution is done with this code:
>>
>> OlapStatement statement = olapConnection.createStatement()
>> statement.executeOlapQuery(query)
>>
>>
>> As for the log, i've added it here so the email is shorter:
>> http://pastebin.com/Y7Z0Hi0b
>>
>> Any ideas are appreciated!
>>
>>
>> On Wed, Nov 13, 2013 at 5:54 PM, Julian Hyde <julianhyde at gmail.com>wrote:
>>
>>> This looks very much like http://jira.pentaho.com/browse/MONDRIAN-1047.
>>> Worth checking whether we have regressed.
>>>
>>> Julian
>>>
>>> _______________________________________________
>>> Mondrian mailing list
>>> Mondrian at pentaho.org
>>> http://lists.pentaho.org/mailman/listinfo/mondrian
>>>
>>>
>>
>>
>> --
>> Hilario Fernandes
>>  _______________________________________________
>> Mondrian mailing list
>> Mondrian at pentaho.org
>> http://lists.pentaho.org/mailman/listinfo/mondrian
>>
>>
>>
>> _______________________________________________
>> Mondrian mailing list
>> Mondrian at pentaho.org
>> http://lists.pentaho.org/mailman/listinfo/mondrian
>>
>>
>
>
> --
> Hilario Fernandes
>



-- 
Hilario Fernandes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20131118/9e030b30/attachment.html 


More information about the Mondrian mailing list