Cognos 8/10 perfomance tuning

I never really understood completely the 'Tuning' settings on dispatchers but I have played with them enough to see how they work and the best combination based on hardware available and want to share those here.

  • For a quad-core process the minimum requirement for RAM I would recommend is 16gb(Cognos 8.3 and higher)- I saw a lot of perfomance improvement in load balancing and scheduling.
  • Set up Processing Capacity to 2 or 1 based on how you have the distributed install and the hardware on each of those machines.
  • High and Low affinity settings should ablways be default values of 1 and 2.
  • Maximum number of processes for the batch report service,report service during peak and non-peak hours can be changed to 8(based on how stable the system is you can increase or decrease if you see unexplained errors or outages)
  • Set the peak time hours according to your business hours and the schedule times
  • Set Queue time limit of the report service (seconds) to 600
  • If you have PPES then under powerplay tab set Maximum Processes to 3 for quad-core processor environment.
  • Setup jobs to delete dump/garbage files once or twice everyday from the following location \bin,\temp,\data
Recycle cognos servers once every week.

Comments

  1. AnonymousJanuary 9, 2012 at 10:56 AM

    Hi Santosh..

    What do you mean by 'recycle congos servers once every week'? I've seen this tern used before. Does it mean restarting the cognos services, or restarting the whole server it is installed on? Anything else involved in this?

    Naveen

    ReplyDelete
  2. Santosh KanthetyJanuary 9, 2012 at 2:44 PM

    I have implemented rebooting of servers as part of recycling cognos services. This will ensure any temp/garbage data or process ID's are released. System reboot can be automated too based on your multiserver architecture, make sure the Cognos service is set to start Automatically.

    ReplyDelete

Post a Comment

Popular posts from this blog

Query Cognos 8/10 Content Store

I came across these queries that can give you good insight on the data and maintenance of Content Store in Cognos 8/10. You can also leverage the CMTools utility that comes with cognos install. ..\Cognos\bin\utilities\cm\CMTools\CMTools.exe select count(*) as "total number of objects" from CMOBJECTS; select c.NAME as "class name", count(*) as "number of objects" from CMOBJECTS o, CMCLASSES cgroup by o.CLASSID, c.CLASSID, C.NAME having c.CLASSID=o.CLASSID order by c.NAME ; select c.NAME as "class name", count(*) as "number of secured objects" from CMOBJECTS o, CMCLASSES c, CMPOLICIES pwhere o.CMID=p.CMID and p.POLICIES is not null group by o.CLASSID, c.CLASSID, C.NAME having c.CLASSID=o.CLASSID order by c.NAME ; select count(*) as "Total number of outputs", SUM(cast(coalesce(DATALENGTH(DATAPROP),0) as bigint)) as "Total size of outputs" from CMDATA where DATAPROP is not NULL; select
Read more

Cognos 8/10 CAMID-User ID/LDAP ID report

I came across this slick KB article on how to get the list of active sessions/users currently on any cognos environment and implemented it. https://www-304.ibm.com/support/docview.wss?uid=swg21385278 After I got this to work the servlet that runs only retrieves the CAMID and the session start time. I started looking around in the content store tables and was able to successfully  retrieve the CAMID and the associated  LDAP ID/User ID. I made this a query subject in the Audit model and joined it to the security groups table and created a report that will give me User ID, CAMID, AD Group, Name , Email. Query : SELECT  distinct ldap_id,     last_name,     first_name,     email,     objid as CAMID     FROM            (SELECT c33.cmid user_id, UPPER (c33.NAME) ldap_id,                   c1.surname last_name, c1.givenname first_name,                   c1.email email, c1.objid              FROM cmobjprops1 c1 LEFT OUTER JOIN cmobjprops33 c33                   ON c33.cmid = c1.cmid          
Read more

Cognos Cross-tab Report Studio Report Ungroup Items to display as list on every line

I recently had a requirement where the user wanted data from an existing Crosstab report to repeat every line so he can use the output in excel for comparison/analysis. It must be an old trick for most of us as I see so many posts around but none of them are straight forward and are hard to follow so i thought why not log it here, besides its only been 10 months since I shared anything. Sorry guys I'm in support right now so no time to explore new tips and tricks or maybe i'm being lazy ;) By default Crosstab groups columns nested based on the position on the cross tab, the trick here is to identify your lowest and highest level of data granularity. Lets pick the below crosstab as an example with PRODUCT, DESCRIPTION, LOCATION as nested dimension columns and default measure of SALES $ across MONTHS.                   Measure: Sales $      MONTH 1 | MONTH 2 .................................... Product | Description | Location |       $10             $20 Cognos groups PRO
Read more