Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Analysis Studio Not Functioning

Status
Not open for further replies.

derdor

MIS
Jul 11, 2006
21
US
We have a peculiar problem... we have Dev, PreProd, and Prod environments. The developers did some updates to a package that consists of a few Report Studio Reports and two cubes... nothing major to anything. Developers develop in Dev-Windows. Then export the package and give to us. We loaded into a Dev-Unix environment and things appear OK. We load into PreProd-Unix (which mirrors Prod-Unix) and Analsys Studio will not run (opening a cube or just opening Analysis Studio itself). We get "No connection to the datasource could be established" on the screen. And the following in the log file "Cannot forward affine request because no node ID was specified. The request is malformed." We have Cognos stumped and we're stumped. We resorted back to the earlier version of the package and it seemd to run fine. We completely reloaded Cognos and the new package from scratch and still have problem. It has to be in the package(?). We've gone over and over the obvious stuff with connections and datasources defined. Anyone see the "affine" errors or any other ideas?
 
interesting problem.. i am not certain how to fix it.

but here is what i would try:
copy all the content from teh webcontent directory (on the gateway box) to webapps/p2p2 (on the app server). now, when you use the dispatcher URI (found in Cognos Connection) you should see cognos connection properly. at this point, we are not going through the web server. see if you run into the same issue.

another (related) thing to try is to ensure that all your server names in cognos configuration are fully qualified (in the environment where you are running into the issue). so instead of just etc, use etc.

as far as i know those asynch errors occur:
1. because of the afore mentioned fully qualified names.
2. because of memory issues.

you can *try* to lower the disptacher memory/sort buffer a little bit as well.
 
I think we found the culprit... not sure if it is a Cognos problem or the way it works yet. We have 2 cubes in the package. One of the cubes seemed to have a problem and did not build in the Unix environment... so this one cube that wasn't available (exist) where it was supposed to, broke all other cubes and analsysis studio in general, from functioning. Once we created the cube (or copied it), the cubes and analsysis studio function OK.
 
hmmm.. as far as i know, you are only supposed to have one cube in one package. why do you have two?
 
Well, that's a good question... I didn't do the development... however, no where can I find from Cognos that says you should only have 1 cube per package, but nor can I find where it says it's OK to have more than 1. I am relatively new to 8 (but have 10+ years of prior Cognos), but it would seem fairly restrictive to only allow 1 cube per package? If I have an area/department with a dozen reports and 5-6 cubes - I need to deploy 5-6 packages? We have been working with Cognos support on this issue, and they have not indicated anything like that. They have logged a bug, but now they are wavering a little now saying it might be working as designed... I'll let you know.
 
well.. why would you have more than one per package? i dont' quite follow.. what advantage could that have? if you try to pull things from many cubes into a crosstab, it will error out..
 
ps.. i don't claim to know a lot about cubes and transformer.. so.. please correct me if i am wrong.
 
The example that Cognos gives is 1 cube in 1 package. I have tried to add a 2nd cube and it doesn't work properly. When I add that cube to a new package it works fine. It would be a shame if you cannot add more than one cube to a package. I was kinda plannin on it. Then you can union them together in the package and really get crazy. Guess that ain't happenin.

CP [cook]
 
In fact you can have as many cubes as you like in a package, but there are some restrictions in how they can be used which may impact your decision to have more than one. If you have more than one cube (or a mixture of cube and query subjects), you cannot bring data from both into a single query. You can bring them in using separate queries in Report Studio, then use use UNIONS or Master/Detail Query relationships betewwn the queries, but this makes the end-user experience far less intuitive.

Just my tuppence...

MF.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top