tag:blogger.com,1999:blog-7751447442026377325.post988553375620754802..comments2023-10-23T22:25:46.368+02:00Comments on Oracle Business Intelligence OBIEE 101: OBIEE cache management.John Minkjanhttp://www.blogger.com/profile/08228393663535024706noreply@blogger.comBlogger21125tag:blogger.com,1999:blog-7751447442026377325.post-2375029623191713972011-08-30T08:10:27.206+02:002011-08-30T08:10:27.206+02:00Hi John,
I have a question for you.
Before that I ...Hi John,<br />I have a question for you.<br />Before that I would like say my background check.<br /><br />In my previous project, we each of us from the DEV TEAM had a separate userid & pwd & access to all SAs.<br />And cache entries were personal.<br />If I get any hit & i want to extract the physical sql , I used to login with another UserID ( Act As ) & fire the same request & could get the physical sql.<br /><br /><br />Now in my personal laptop , I have installed OBIEE 10g. I seeded a cache for 1 small report using this command- <br />NQCMD -d AnalyticsWeb -u Naresh -p Naresh -s C:\SeedCache.txt<br /><br />Now when I login using another user say Suresh, I can see the cache hit in the NQQuery.log.<br /><br />why the cache is getting shared ?Nachikethttps://www.blogger.com/profile/00817531303871825351noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-67381818527880606382011-06-28T23:19:28.455+02:002011-06-28T23:19:28.455+02:00Here's something I've not read anywhere. ...Here's something I've not read anywhere. OBIEE will fail to give you a cache hit if your report has fewer columns than the # of columns in the caching report. In other words, if the report has Columns A and B and you have a cache entry with Columns A, B, and C this will not result in a hit. This is contrary to the doco and contrary to intution. I've got nothing else special going on in my testing, running OBIEE 10.1.3.4.1 against an Oracle database. IMO this almost makes caching useless and I can't believe I haven't read this anywhere else (have not search the Oracle SRs yet).David Andersennoreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-67431231432258646022010-07-19T18:25:30.817+02:002010-07-19T18:25:30.817+02:00We are thinking of using "in memory" dri...We are thinking of using "in memory" drive to improve the cache performance. However, this is a new area, and could someone suggest good software/hardware for "in memory" drive? Thanks.Changhttps://www.blogger.com/profile/13388164327944806300noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-41142679307989709912010-03-29T11:36:14.327+02:002010-03-29T11:36:14.327+02:00Hi John,
first of all thank you for your very usef...Hi John,<br />first of all thank you for your very useful guides and suggestions. If I set the size of cache to 2 GB, BI Server does not cache anything because it allocates 0 MB to the size! But if I set the size to 2000 MB (instead of 2 GB) it regularly works and cache management "sees" 2000 MB of available space. This happens on Windows. Do you have feedback similar to this one? Thank you!Gaetano Liscohttps://www.blogger.com/profile/10207559854213935144noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-18916206148058081962010-03-17T14:58:07.336+01:002010-03-17T14:58:07.336+01:00Good arcticle. One quick question.
If a report is...Good arcticle. One quick question.<br /><br />If a report is a pre-scheduled report with number of drill-through capabilities how does the caching work. Does the report and the drill through paths get cached.Aniruddhahttps://www.blogger.com/profile/15242378202457140681noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-25891138962884619762010-01-05T18:48:45.517+01:002010-01-05T18:48:45.517+01:00Hi I am new to OBIEE and lookging for detailed doc...Hi I am new to OBIEE and lookging for detailed document about how the request get processedin OBIEEAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-46941622583585708572009-04-23T15:39:00.000+02:002009-04-23T15:39:00.000+02:00@Rits,
Sadly no.... I have a great lack of time a...@Rits,<br /><br />Sadly no.... I have a great lack of time at the moment. And to investigate this would take me a couple of hour to set up.<br /><br />regardsJohn Minkjanhttps://www.blogger.com/profile/08228393663535024706noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-72905476113773933412009-04-23T12:53:00.000+02:002009-04-23T12:53:00.000+02:00Hi John,
Did you get chance to test / validate MS...Hi John,<br /><br />Did you get chance to test / validate MSP's comment? <br /><br />Can you please share your finding?Riteshhttps://www.blogger.com/profile/04870525811305969141noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-76400444357234805992009-02-27T11:47:00.000+01:002009-02-27T11:47:00.000+01:00Hi Msp,I'll get back to you ASAP. I will set up a ...Hi Msp,<BR/><BR/>I'll get back to you ASAP. I will set up a test enverioment for this behauvior first.<BR/><BR/>regards<BR/><BR/>JohnJohn Minkjanhttps://www.blogger.com/profile/08228393663535024706noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-27266420850761998702009-02-18T15:28:00.000+01:002009-02-18T15:28:00.000+01:00John,Thanks for explaining the workings of OBIEE c...John,<BR/>Thanks for explaining the workings of OBIEE cache in such great detail. Could u please clarify the following -<BR/>I was running some tests on my local system and I observed that a user’s request only will only get a cache hit on that same user’s cache.<BR/><BR/>We are not using VPD. As per the Admin guide -<BR/>“Initializing Cache Entries for User Ids<BR/>To initialize cache entries for user Ids, the Connection Pool needs to be set up for shared login ..If the shared login is disabled and a user specific database login is used, cache will be shared.”<BR/><BR/>Does the above line from the Admin guide explain the behavior I see or am I missing something ?<BR/>I as this because, based on your article, and some other postings, it seems that if no row level security (VPD) is implemented, the cache should be shared among users. But that is not happening and I would like to understand why.<BR/><BR/>ThanksAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-17904888898916015202009-01-05T19:53:00.000+01:002009-01-05T19:53:00.000+01:00Hi Christian,You are totally right, that sentence ...Hi Christian,<BR/><BR/>You are totally right, that sentence went wrong in translation. I have added the correct syntax.<BR/><BR/>regards JohnJohn Minkjanhttps://www.blogger.com/profile/08228393663535024706noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-26547049388644909852009-01-05T14:24:00.000+01:002009-01-05T14:24:00.000+01:00I don't think this is correct: "The cache entries ...I don't think this is correct: "The cache entries are personal, so a different user executing the same request will lead to a new cache entry"<BR/><BR/>I can see why the query will be different if you are using row level secrurity. But if the query is the same them OBIEE uses the cache for a different user. I have tested and I am quite sure about it.Turribeach.com.arhttps://www.blogger.com/profile/03199234875197777117noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-5997947029613271122008-11-26T07:51:00.000+01:002008-11-26T07:51:00.000+01:00Hi John,First of all I'm not a fan of using the OB...Hi John,<BR/><BR/>First of all I'm not a fan of using the OBIEE cache. It's usually a sign of the DWH not complying with user needs. I see the OBIEE cache more as a speed up if the same user / group is a asking the (partly) same question within the same ETL window. From my experiences there is often more gain in tunning the DWH to the users need.<BR/><BR/>regards<BR/><BR/>JohnJohn Minkjanhttps://www.blogger.com/profile/08228393663535024706noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-11747841975451180122008-11-24T22:48:00.000+01:002008-11-24T22:48:00.000+01:00I'm baffled why I don't get a cache hit when I run...I'm baffled why I don't get a cache hit when I run a request from Answers using the exact same criteria as was used to seed the cache. The cache was created by a different user via an iBot answers request. We are not using row level security in this case, although everyone has their own login, authenticated via our own home grown portal security. Why is cacheing so often discussed as a nice feature if you can't hit against anyone elses cache? We have too many users to possibly run cache seeding requests for all of them. Am I missing something here?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-25912646028269638492008-11-22T21:44:00.000+01:002008-11-22T21:44:00.000+01:00Hi Raghev,Please share your toughts on the subject...Hi Raghev,<BR/><BR/>Please share your toughts on the subject.<BR/><BR/>regards JohnJohn Minkjanhttps://www.blogger.com/profile/08228393663535024706noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-69252179381949230162008-11-22T09:59:00.000+01:002008-11-22T09:59:00.000+01:00Hi,Can u explain what are the Qualified misses and...Hi,<BR/><BR/>Can u explain what are the Qualified misses and unqualified misses in the cache entries...<BR/><BR/>If Not Let me try...Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-43162568592702283812008-11-10T21:20:00.000+01:002008-11-10T21:20:00.000+01:00I've just opened an SR with Oracle because I'm try...I've just opened an SR with Oracle because I'm trying to run requests through iBots to seed the cache but have found that our session variables from init blocks are not being populated. As we've employed these session variables in our LTS it is preventing the requests from running via iBots and/or Job Manager. <BR/>Great blog, very helpful to many of us.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-53711404427188331052008-11-10T20:29:00.000+01:002008-11-10T20:29:00.000+01:00Hi John,Yeah, that should work. The thing is that ...Hi John,<BR/><BR/>Yeah, that should work. The thing is that OBIEE is kind picky when a a session variable is in the query. Why don't you seed the cache with a general query and then run a couple of querys for different users. After that you check the NQquery log if you it picked up the cache or not. (Set the logging level for the users temporary to 2 or higher)John Minkjanhttps://www.blogger.com/profile/08228393663535024706noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-32337530790957399382008-11-10T18:07:00.000+01:002008-11-10T18:07:00.000+01:00We have row level security where a Logical Table S...We have row level security where a Logical Table Source restricts data selected based on a list of account numbers from a Session Variable. The USER variable is not used in the LTS "where" clause. Wouldn't any user's query selecting against an account number that is in the other user's cache get a cache hit? <BR/>(note that the init block which populates the account# session variable does refer to the USER variable in its selection of accounts from a security table).<BR/>Thanks in advance,<BR/>JohnAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-83226004757239140492008-11-07T21:47:00.000+01:002008-11-07T21:47:00.000+01:00Hi John,That depends on the datasecurity model you...Hi John,<BR/><BR/>That depends on the datasecurity model you are using. With user / rowlevel security you will get a need a seperate seed for every user. If it your security is on "group level" you need a seed for each group. If all users are allowed to see all data, one entry is enough.<BR/><BR/>regards <BR/><BR/>JohnJohn Minkjanhttps://www.blogger.com/profile/08228393663535024706noreply@blogger.comtag:blogger.com,1999:blog-7751447442026377325.post-26485655435282827862008-11-07T21:07:00.000+01:002008-11-07T21:07:00.000+01:00You said that cache entries are personal, so a dif...You said that cache entries are personal, so a different user executing the same request will lead to a new cache entry. Would that mean that if I seed my cache I have to run a cache seeding request once for each possible user of my data?Anonymousnoreply@blogger.com