cancel
Showing results for 
Search instead for 
Did you mean: 

Issue with WEBI reports based on Multi Source Universe

Former Member
0 Kudos

Hi Folks,


We are facing an issue with the WEBI report based on the multi-sourced

universe.


Environment - BO 4.1 SP5

Data source - 2 Oracle 11.2 sources


When we select the prompt values and refresh the report it throws the

error as

Database error: [Data Federator Driver]

Cannot Reach Server. Reason: attempt to establish connection failed:

java.net.ConnectException: Connection refused: connect. (IES10901)

(Error: INF)


rgds,

Prakash S

Accepted Solutions (0)

Answers (3)

Answers (3)

Former Member
0 Kudos

Hi,

Refer SAP Note: 1901933

Cause

The connection to the host is not available. Possible causes are:

  • Incorrect host or port entry in the ODBC.ini
  • Firewall blocking host/port destination
  • Client is unable to resolve hostname of the database host

Resolution

  • Changes need to be made to the areas defined by the causes above. The ping command can be used to determine if a host can be reached by hostname or IP.
  • If unable to connect via host name or IP then the issue is likely with a firewall.
  • If unable to connect by host name but can connect by IP the issue is with name resolution.
  • If both can be pinged then the issue is likely an incorrect entry or syntax in the ODBC.ini
Former Member
0 Kudos

I compared my DF service for DEV with the PROD one and found some differences in the common services:

Prod had:

  • Adaptive Connectivity Service
  • Data Federation Service
  • Tracelog Service

Whereas DEV had:

  • Adaptive Connectivity Service
  • Custome Data Access Service
  • DSL Bridge Service
  • Data Federation Service
  • Document Recovery Service
  • Excel Data Access Service
  • Rebean Service
  • Security Token Service
  • Tracelog Service
  • Visualisation Service
  • WebIntelligence Monitoring Service

The same person set up both so not sure why they aren't the same..

As a test I stopped and disabled the prod one and created a new one with the same services as the DEV one enabled it and set it to start with the SIA and rebooted the prod server.

It seems to have fixed the issue, but I'm not sure that all those common services are required ?

The other thing is that the original prod service had some monitoring conditions set up with the Caution Rule being

Server.APS.DataAccess$'Server Running State'==1 || Server.APS.DataAccess$'Server Running State'==2 || Server.APS.DataAccess$'Server Running State'==4 || Server.APS.DataAccess$'Server Running State'==6 || Server.APS.DataAccess$'Server Running State'==7

and the Danger rule being

Server.APS.DataAccess$'Server Running State'==0 || Server.APS.DataAccess$'Server Running State'==5

And If log in to webi as Administrator I can see a whole bunch of messages saying these rules have been crossed but only seem to have started showing for the last 24 hours even this issue has been randomly happening for weeks and I spent most of 2nd November trying to wor it out and getting numerous crashes more than 24 hours ago.

The new service I created doesn't have any monitoring so it's not set up by default.

Former Member
0 Kudos

Just further the person who set up the APS in the different environment said it was because Prod has 32GB memory whereas DEV only has 8GB and suggested changing it back even though it reintroduces the error.

In regards to pinging, I can ping both server and host name and I'm using OLEDB connection so nothing to do with ODBC.

denise_dunker
Explorer
0 Kudos

We had an issue earlier this year.  The Data Federation service would run fine for a while, then crash, creating a dump file.  After the crash, the data federation aps  would have to be restarted.  Database is SQL Server 2012.  We found a Microsoft article about a memory leak in OLEDB.  We changed all our connections to ODBC and have not had that problem since. 

former_member219634
Active Participant
0 Kudos

Dear Denise

  Could you please give us the link for Microsoft article about lead in OLEDB?

Best Regards

Ines

denise_dunker
Explorer
0 Kudos

I can't seem to find the original one(been a long time); and possibly Microsoft has released a patch to correct.  Here looks to be similar to what i found.. https://support.microsoft.com/en-us/kb/2881661  .  We  could see the following   in the Stack dump.

--------------  N a t i v e   C r a s h i n f o  ---------------

Stack:

    SP                 PC                 RetPC              Symbol

0: 0x000000002a9be110 0x0000000053fb1391 0x0000000053fd4f37 dbd_wsqloledb::????

Former Member
0 Kudos

Same issue, randomly, using SQL Server back end.

former_member193866
Active Participant
0 Kudos

Hi Sivaraj,

Check servers once in CMC

Thanks

Riaz

Former Member
0 Kudos

Hi Riaz,

The servers seems to be OK. However, whenever we have this issue, we see the date modified date was of the APS that has Data Federation service has the different last modified date i.e. recent last modified date

rgds,

Prakash S

Former Member
0 Kudos

I purely by chance noticed that what seems to be happening is the Data Federation service is crashing and restarting.

I found this because I worked out if I have the crash and log out of Webi and back in the report seems to work for a few refreshes before crashing.

Once crashed 5 times though I was getting some error about not be authorized to return data from the universe.

I logged into the CMC and by chance notice that the DF service had stopped - with the explanation being it had crashed 5 times within and hour.

I restarted it, switched to Webi to run the report and when it crashed I happened to flick to CMC and saw the service was initializing.

I've tried stripping back my report query by query to see if it's a specific data provider with no luck, even went through the universe and gave unique names to each object as I think I saw that suggested somewhere - still no luck.

Because of the way our environment is set up we have to have 3 sets of connections and 3 .DFX and repoint the universe to the appropriate DFX before promotion.

Because I've been playing around with all of the various components trying to fix this issue, our DEV Webi environment is currently using the PROD connections ad universe.

I happened to notice last night that it doesn't seem to crash in DEV !!! i.e. I ran the report about 20 times with various prompt values and no issue.

So, now I have to work out what the difference is between the two environments - needle in a haystack - could be any setting any where or something in the promotion - meanwhile my boss is suggesting we 'give up' on multi source universe and try and rework the report using two single source universes.