cancel
Showing results for 
Search instead for 
Did you mean: 

List of Values (dynamic parameter) times out once and then works.

Former Member
0 Kudos

Hello. First time posting here. I'm having a strange (yet very predictable) issue with Crystals Reports List of Values. The List of Values is used on a dynamic cascading prompt and is published to the BO XI R2 repository. The prompt/list works fine when using Crystal Reports thick client, but when using the Crystal Reports viewer from InfoView the following happens everytime:

1. Any Report that uses the Parameter/LOV is launched.

2. After about 1-2 minutes, the viewer returns a "Timeout" error message.

3. The report is launched a 2nd time.

4. The parameters/LOVs display (takes about 8 seconds) and the report runs fine.

Then anytime a report is run within a certain period of time it works ok. But if no one runs a report for an extended period of time (i'm not sure how long, but it's on the order of maybe 30 mins.. to an hour), then it starts all over again. In addition, if I restart the services/server this same thing will immediately occur. (time out once, but work after that for a period of time).

I initially did not schedule the LOV to get generated on the server and left it so that it's done in real-time when the report ran. But then when I ran into this error, I tried scheduling it every 10 minutes or so, but the error continues to happen.

I also restarted the server/etc to no avail.

I saw another post on here that was kind of similar and that suggested that starting the LOV services as a domain user (that had admin rights on the BO server) helped him. I tried this as well and restarted everything but that didn't fix my issue.

Any help would be greatly appreciated!

Scott

Accepted Solutions (0)

Answers (1)

Answers (1)

Former Member
0 Kudos

Hi Scott.

How long does the query to generate the LOV take? It sounds like the db is caching the results and therfore dramatically reducing the time to refresh. You might consider extending the -requestTimeout on the list of value and RAS services, to ensure they cover the timeframe needed to refresh the LOV.

Also in the Business View manager, what do you see as the List Of Value status?

Regards

Alan

Former Member
0 Kudos

Hi Alan. Thx for the quick reply. Just fyi that i'm kind of a newb regarding Crystal Enterprise and all of that. I used Business Objects Universe and Desktop Intelligence quite a bit, but all of the Crystal side of things (Business Views, LoV, Parameters, etc) is kind of new to me.

The backend database is SQL 2000 btw. The query that builds the dynamic cascading parameter is pretty quick. It takes less then 2 seconds to get all of the results back using Query Analyzer over a WAN connection. (the Crystal stuff all runs on the Local Area Network btw).

In BV Manager, the status was at first something regarding "no schedule is defined, so it's dynamically created..blah blah". (don't remember the exact wording).

After enabling the schedule the other day, the status currently looks something like:

Status of List of Values Instance: Success

Creation Time: 2/3/2009 2:54:07 PM

Start Time: 2/3/2009 3:04:44 PM

End Time: 2/3/2009 3:04:46 PM

Server Used: SFBO01.ListOfValuesJobServer

A quick question: I've seen other forum posts that indicate that there is a way to have the ListOfValues service create a more detailed trace/log. I've done some quick searching around and haven't yet found how to enable that. Do you happen to know offhand how I could turn that on so I can capture more detail on what is happening?

Thanks

Former Member
0 Kudos

Hi Scott,

you need to add -trace to the end of the server command line. In CMC go to the servers section. From there you should be able to select the server your are interested in List of Values and RAS and in the properties it is possible to edit the command line

Regards

Alan

Former Member
0 Kudos

Thanks Alan. Those settings definitely turned on tracing. Lots of info dumped in those trace files and I don't see anything that seems to stick out (Of course I don't completely understand how to interpret it...).

Very strange... I restart the services and everytime, the same exact issue. Report times out once (about about 1-2 minutes) and then it works fine for a period of time.

It's as if the 1st go is doing something to set things up and whatever that is, it's taking a very long time... but then once that is setup (again, whatever that is), it works ok.

I have a feeling that because we are a "BO/Universe" shop, I probably don't have all of those Crystal job services setup propery.

Former Member
0 Kudos

Hi Scott,

This stuff should be working out of the box, there is no special settings to get it going.

One question, have you tried to schedule the report from the CMC. There may be an InfoView timeout rather than a DB timeout.

Regards

Alan

Former Member
0 Kudos

Well no, but it's not the reports that are timing out (I don't think), it's the parameters/prompts. I can't schedule the reports because there are hundreds of options/selections in the prompts that make the report different for anyone running it. I have created other dynamic prompts for Crystal/BV that work fine. The only difference here that I can think of is that these new prompts are a.) cascading and b.) contain more data.

Just to give you an idea of the size of the data, there are about a total of 1000 records returned by the query (slightly less then 1000). The top level dropdown contains 530 items and then there are between 1 and 5 values in the 2nd dropdown depending on what you pick in the first drop down.

Scott

Former Member
0 Kudos

Hi Scott,

So this is happening when you refresh the reports on demand? I understand that you cannot schedule the reports in the long term, but you could try it now to see if the LOV displays correctly when scheduling.

You might also try viewing the report from the CMC, as I said this might be an infoview specific setting rather than anything wrong with the underlying report/lov

Regards

Alan

Former Member
0 Kudos

Thank you. I never run reports from the CMC and that's a good tip.

Here is what happens when I "Preview" the report in the CMC:

1. I find the report and click on "Preview"

2. A new browser window opens and after a couple minutes, I get a runtime error from the web server (the error is suppressed so I'm not exactly sure what the error text is).

3. I close that window and click "Preview" again.

4. This time, the parameters come up fine and the report runs fine.

So, it's basically the identical situation whether it runs from the CMC or Infoview. The first time the report is launched, something times-out creating the LOVs/parameters. Then, if launched again, it works fine. This now runs fine for 30 minutes or so and then the whole thing starts over again.

Former Member
0 Kudos

Hi Scott,

Can you post the actual text of the Timeout error you receive from InfoView

Regards

Alan

Former Member
0 Kudos

The actual text is:

"An error has occurred: Request timed out."

Former Member
0 Kudos

Hi Scott,

One other place you could try is the Tomcat logs, after that I would raise an issue with BO, passing the trace log details. If you get a response can you post it here.

Regards

Alan

Former Member
0 Kudos

Just wanted to update this.

Because I use Business Objects OEM, I didn't have direct support with SAP/Business Objects. However, because of this issue (and other issues) we purchased Business Objects Edge 3.0.

After getting BOE 3 installed/configured, I migrated those reports and underlying Business Views over.

And guess what? The problem still occurred on the new platform.

I then created a support incident and the person that initially tried to help me couldn't figure out what was going on.

However, during that time, I did try re-creating the report using a newer version of Crystal (Crystal 2008) and the problem went away. (note, I didn't recreate the business views.. just the reports).

So bottom line:

I created all reports from scratch instead of using the reports that I migrated from BO XI R2/CR XI and no longer had the problem.

The techs never figured out the exact cause of the issue.

So far, my experience with their support has been the same on all incidents. They don't seem to be trained (at least at the 1st level) for actually using the debug/logs/etc to help quickly pinpoint/analyze issues. They do basically like I do (the end user) and smartly poke around and do trail/error. I have yet to get a support tech that knew the exact nature of my issue and could solve it without this kind of trial/error approach.