[Mondrian] StackOverflowError

Julian Hyde jhyde at pentaho.com
Fri Jan 6 15:52:16 EST 2012


Thanks. It is a Mondrian issue, then. I have re-opened the case.

The issues you are having with Workbench are a separate issue. Log a jira case for them if you like.

Paul: It would have helped if you guys had propagated the Mondrian stack back to the client or at least into your log.

Julian

On Jan 6, 2012, at 6:39 AM, Kristopher Kiger wrote:

A quick follow up to what I sent last night.  I put workbench into debug mode and loaded the schema file, without the recursive measure in the virtual cube, successfully.  A success confirmation was displayed when connecting to it via workbench’s mdx tool.  I then added the recursive measure reference back into my schema and tried to connect.  I observed the same behavior as stated last night.  No success or failure dialog was brought up when trying to connect with the mdx tool.  I wrote a quick query to make sure it did not load the schema file and was given an error that it did not know what that member was.  I checked the log file and found this exception:  http://pastebin.com/inkNtG3D

Kris

From: mondrian-bounces at pentaho.org<mailto:mondrian-bounces at pentaho.org> [mailto:mondrian-bounces at pentaho.org] On Behalf Of Kristopher Kiger
Sent: Thursday, January 05, 2012 9:34 PM
To: Mondrian developer mailing list
Subject: Re: [Mondrian] StackOverflowError

I can say that I was unable to get the schema to load successfully into workbench.  It did not spit out an exception or anything.  Generally, when loading a new schema version, I will try to connect to it via the mdx query interface within workbench to further validate the schema file.  When it connects/validates successfully, it will spit out a generic "connection successful".  However, when I include the virtual cube with the specified measure, it does no such thing (there's no explicit failure, either).  I will see if I can increase workbench's logging verbosity tomorrow morning.  Perhaps there's something I'm not seeing at the current level.

In the context of Julian's latest question, the exception I listed was one produced using saiku server (standalone) 2.1.

Kris
________________________________
From: mondrian-bounces at pentaho.org<mailto:mondrian-bounces at pentaho.org> [mondrian-bounces at pentaho.org<mailto:mondrian-bounces at pentaho.org>] on behalf of Paul Stoellberger [p.stoellberger at gmail.com<mailto:p.stoellberger at gmail.com>]
Sent: Thursday, January 05, 2012 9:24 PM
To: Mondrian developer mailing list
Cc: Mondrian developer mailing list
Subject: Re: [Mondrian] StackOverflowError
i cant spot anything saiku core related in that log. is there anything missing in that log?
an you execute a query with that measure in schema workbench?

i know that we experience some weird issues with virtual cubes but the info i have is not sufficient enough to narrow it down to our or mondrian code.


-paul



Am 06.01.2012 um 03:16 schrieb Julian Hyde <jhyde at pentaho.com<mailto:jhyde at pentaho.com>>:
I've commented in that bug that it seems to be a Saiku issue.

Paul, can you give me your opinion?

Julian

On Jan 5, 2012, at 8:15 AM, Kristopher Kiger wrote:


Julian,

I posted a bug report on jira here:  http://jira.pentaho.com/browse/MONDRIAN-1061

I was able to reproduce the issue with a modified version of the foodmart schema.  A link to that is provided in the bug report.

Thanks for your help!

Kris

From: mondrian-bounces at pentaho.org<mailto:mondrian-bounces at pentaho.org> [mailto:mondrian-bounces at pentaho.org] On Behalf Of Julian Hyde
Sent: Wednesday, January 04, 2012 3:38 PM
To: Mondrian developer mailing list
Subject: Re: [Mondrian] StackOverflowError

Doesn't sound familiar. Google for mondrian stackoverflowerror and see if any previous issues match. If not, please log a jira case with the full stack.

If you can repro the issue on a variant of the foodmart schema, that would be awesome.

Julian

On Jan 4, 2012, at 12:12 PM, Kristopher Kiger wrote:

I'm working with a virtual cube, which makes reference to an inventory cube.  The inventory cube has a recursive calculated member within it.  Alone, the inventory cube and recursive member work great.  However, when I incorporate that member within a virtual cube by using a VirtualCubeMember, I get a java.lang.StackOverflowError.  If I create a new calculated member within the virtual cube and paste that same recursive code it loads into mondrian with no problems.  It’s only a problem when I try to include the recursive member as a VirtualCubeMeasure.  Has anyone run into a similar issue?  If more detail is needed, let me know.  I’ll do my best to supply it :)

Cheers,
Kris

_______________________________________________
Mondrian mailing list
Mondrian at pentaho.org<mailto:Mondrian at pentaho.org>
http://lists.pentaho.org/mailman/listinfo/mondrian

_______________________________________________
Mondrian mailing list
Mondrian at pentaho.org<mailto:Mondrian at pentaho.org>
http://lists.pentaho.org/mailman/listinfo/mondrian

_______________________________________________
Mondrian mailing list
Mondrian at pentaho.org<mailto:Mondrian at pentaho.org>
http://lists.pentaho.org/mailman/listinfo/mondrian
_______________________________________________
Mondrian mailing list
Mondrian at pentaho.org<mailto:Mondrian at pentaho.org>
http://lists.pentaho.org/mailman/listinfo/mondrian

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20120106/f4a524bd/attachment-0001.html 


More information about the Mondrian mailing list