We ran into multiple issues because of high data sets.1. Cognos reports
created tomcat memory problems trying to fit all the queried data into its
memory.
2. Multilevel where used reports on Manufacturer parts caused system crashes
because of memory issues.
3. Ran into full list issue while displaying tables.
4. Method Server crashes when exporting search results for large data sets.
All these issues occured even after fine tuning Oracle database , tomcat and
MethodServer. We ended up setting the max query limits to 130K as suggested
by PTC.
We run into minor annoyances because of these limits.
1. Now the Partslink classification update doesn't work with the query limit
set to 130K. We have to reset these properties to -1 to run the
classification update. We are running this tool only during the production
downtime.
2. In some cases with large ProE assemblies with family table data, checkin
and checkout functionality fails after hitting this limit. Actual number of
components in the assembly are far less than 130K. We are still debugging
this issue.
Regards,
Prathap
On Tue, Aug 25, 2009 at 10:09 AM, Lockwood,Mike,IRVINE,R&D <
mike.lockwood@alconlabs.com> wrote:
> Another in a long line of functions in Windchill not being planned for
> scaling to the sizes that users need to deal with for normal work...
>
> PTC absolutely needs to do something more systematic to address how the
> system behaves with large data sets.
>
> No ideas to offer on this one - let us know what you find out.
>
