We have been seeing issue with Search in BI Launchpad, user ends with
the above error message while searching. This happening only on the
Launchpad, searching on CMC working. I also noticed in sometime the Search results ‘No results found in the search’ in BI Launchpad.
Error message-
I have also referred the SAP Note- 1856714 , it has mentioned that this
is a defect. However,it has worked earlier in our environment and the
issue started recently.
I have also performed the below steps and this has not resolved the
problem.
I’m not sure if it’s 100% related to your error message, but the new search engine in R4 has some bugs that are both known and unknown by SAP. (Unknown in that they cannot find the cause as of yet).
The CMC search is the same search that was in XIR3 and therefore is unaffected by the new search platform.
Most of our search issues are resolved by rebuilding the index. This was suggested as a workaround/fix by SAP, and we do this weekly… so far our errors and issues have been resolved by simply rebuilding our index each Friday.
But search is work strange or I don’t know all of it.
I try to find some report.wid but I can’t. I CMC I have:
Crawling Frequency - Scheduled crawling
Level of indexing - Full Content
Content Types - Web Intelligence
and I can’t find my report. When I use RichClient and try import report from server I can find my report in search windows which appears during import.
However if search isn’t returning reports, but the CMC is… then most likely your index is either a) not completely rebuilt, or b) corrupted as my other post indicated happens.
My platform application settings are a bit different:
Crawling Frequency = Continuous Crawling (I don’t like the idea that items are only indexed based on a schedule).
Level = Platform Metadata (I rarely put anything of note to search by in any document’s metadata, so no need to index it.)
Content Types: Limited to Crystal, Webi, universe and Workspace. We don’t have any other types in our environment currently.
To determine if your system is ‘done’, the only true method is to look at the “Number of indexed documents” and record the number. Come back 5-10 minutes later and see if it increases. Keep in mind that unless you have a quiet system (such as middle of night) it might increase as users add documents during normal business practices. If it stops moving, then you know indexing is done.
If you know it’s done and still missing… you can try to rebuild… My steps are a bit extreme, but they do seem to keep my system up and running weekly.
Enter the Platform Search Application properties and click "Stop Indexing. "
Save and Close
Wait about 10-15 minutes to ensure all index routines have stopped.
Go back into the application properties and place a checkbox in the “Rebuild Index” box.
Click “Start Indexing”
click Save and Close.
After that you can continue to monitor it by opening up properties and noting the documents indexed. It will not refresh so you have to close and reopen it again to see the change.
Also I had one instance where it never started indexing (after about 30 minutes) So I simply restarted the index server within the CMC and that seemed to have started things off again. Normally that step is not necessary.
Logged in as the Enterprise Administrator… (After several users complained they receive the message)
Environment: BI4.0 SP06
In BI Launchpad:
An error occurred while searching: A com.crystaldecisions.enterprise.ocaframework.IManagedService$ManagedLogoffException occurred; original exception message Unable to reconnect to the CMS UKBOB03.kgfruits.local:6400. The session has been logged off or has expired. (FWM 01002)
The above message appears immediately after entering the search term and hitting search, I’ve checked the usual timeouts etc and these are all fine.
The Platform Search Service has been split out into it’s own dedicated APS, so nothing else interferes with it.
The really odd thing is… Performing the same search in the Central Management Console brings back results!?!
Thinking it may be something permissions related, I checked the application permissions, can’t see anything wrong there and there’s nothing obvious to check really…
In the end, I restarted the Platform Search APS and waited for it to re-index, this resolved the issue… But If anyone knows what could be causing it, I’d be delighted to hear suggestions.
I’ve not seen that error myself, but there is nothing “odd” at all with what you saw.
The CMC search and the Launchpad search are two totally different sets of code and applications within BO. I don’t know if they share the same permission structure at all, but I do know that the code they run on is totally independent. (And for reference the Query Builder uses the same CMC method)
The CMC is the old search method from XIR3, while the Launchpad is a whole new system built for R4. Eventually I suspect that the CMC search will be replaced as well once they work all the bugs out.
Lord helps us all when they do replace the CMC search with this buggy code that they are using in BI Launchpad right now.
Just wanted to say thanks for the idea of rebuilding the index. It helped me get the search to start working again, although it will take a while to get through the 10K+ objects the client has in their environment.
Actually, the search in CMC does not appear to use the Platform Search results. Too bad, because in 4.1 SP1 Patch 3, the search in the CMC for things like Personal Folders/Inboxes, etc. is VERY SLOW
I too am having fun and games with Platform Search in BI4.1 (SP2 Patch 3).
Is there a consensus on Scheduled vs Continuous? The situation we have is that Continuous was using and continuing to use a lot of disk space. So following investigations and advice I toned down what was indexed and changed from continuous to scheduled. The problem with this is that post completion of a scheduled crawl the index works for a while and then starts producing the error at the top …
A rebuild of the index resolves this but why does this happen to a successfully built index?
I am now testing continuous crawling again with the toned down “what to index” settings.
The other question I have is regarding the time setting for scheduled crawling. When set high does the index building always take the full time? I.e. when it completes the index does it switch to continuous crawling until the time setting is reached and then stop? Or are my rebuilds never actually completing? As then always seem to run for the time I set (even 3600 minute for 5000 odd documents).
Small update from me. Since setting the crawl to be continuous, level of indexing to just Platform Metadata and removing Universe from Content Types it’s been fine. No excessive disk space usage, it picks up new reports and more importantly it’s remained operational.
(BOB member since 2006-02-08)
(BOB member since 2002-08-19)
(BOB member since 2011-03-03)
(BOB member since 2007-07-04)
(BOB member since 2003-10-16)