Essbase Fragmentation Question - Average Fragmentation Quotient

According to the DBA guide, if we are using the 'Average Fragmentation Quotient' to evaluate whether the level likely will cause performance problems. We should defrag if the quotient is above 30%. (our db is over 2G)
How do I read the below DBSTATS reults for that db? is 'Average Fragmentation Quotient' 43.797% or .437977 %? I've had differant answers from consultants on this.
Average Cluster Ratio: 1
Average Fragmentation Quotient: 0.4379771
Free Space Recovery Needed: Yes
Estimated Bytes of Recoverable Free Space
Thank you!

My guess that it is saying 43.79%, which suggests that there is fragmentation, however your Average cluster ratio is 1, indicating no fragmentation....
when I say guess, I mean my understanding.
Edited by: name_and_address_withheld on Jan 10, 2013 7:45 AM

Similar Messages

Essbase MDX question

Hi there
I am at a client where I have to do comparative store analysis. Basically with MDX I need to do the following:
When doing comparable store analysis on a daily level in my Time dimension a store must have traded a year ago on the same comparable day as per the company's trading calendar.
When doing comparable store analysis on a weekly level a store must have traded a year ago for the full same comparable previous year's week (i.e. all the days in the week) as per company's trading calendar.
If the criteria is not met #missing must be returned else the day or week value.
On day level my MDX works fine:
CASE
WHEN
IsLevel([Entity].CurrentMember, 0) AND
IsLevel([Time].CurrentMember, 0) AND
NOT IsEmpty([Periodicity].[LY])
THEN
([LY], [Act], [Time].CurrentMember, [Entity].CurrentMember, [Account].CurrentMember)
END
The question is on a week level how do I check if a store traded on every single day of the same week last year. Getting the week is no problem, but checking every child of that week.
Just another point on this: I need to do this on month, quarter etc levels as well, so hard coding to check from day 1 to 7 will not suffice.
Thanks
Johan

You are correct, currently Essbase does not support updates via MDX.

Essbase Storage Questions

Hi,
I would like to know during calculation how essbase read and write data to disks,
whether it use synchronous IO or asynchronous IO.
And also the block size of data written to the disk.
I know in BSO, we can calculate Essbase block size, however
as this question is coming from storage / infrastructure people,
what they mean probably is for each write to disk,
what's the size of data / block written to the disk? How do we measure that?
This question come up as we would like to tune essbase on Solaris with
zfs file system. Any feedback is appreciated.
Thanks,
Lian

Hi Lian,
You can monitor these metrics in real time using Accelatis.
The tool can provide load generation or profile calc scripts and as it does,
you can watch the effect your activity has the values of these metrics
being plotted on a graph that resembles an EKG.
This information is stored for play back later, and for reporting and analysis.
In addition to standard metrics, Accelatis can also track disk usage,
cache utilization and performance, and many other elements that
affect performance.
Regards,
Robb Salzmann

Essbase export question

I have the strangest error.
I am trying to export Level 0 and import in our new enviornement.
Versions for target and source: 11.1.2.2
My method of export:*
Right click [database name] --> Export
Export to file [ name.txt]
Export option: Level0 data blocks
Export in column format
Expected output.
Header: Begbalance, jan -------> Dec, Period
This output
Header: "HSP_InputValue" "HSP_InputCurrency" "HSP_Rate_USD" "HSP_Rate_RMB" "HSP_Rate_CNY" "HSP_Rate_PLN" "HSP_Rate_EUR" "HSP_Rate_GBP" "HSP_Rate_MXP"
Now I have exported and imported other applications successfully and all of them include HSP_Rate dimension.
other info if you need:
These are EPMA history applications of from 11.1.1.3.
I am creating new BSO applications in our OOD 11.1.2.2 and importing the data (yes i am not migrating the planning app itself, just migrating the data into this bso)
please let me know what you think.
I have even tried restructuring the outline since HSP was the first dimension, i dont know why that was the first. But either way I relocated it in essbase. (did not do it in planning because it was deleted long before. Only essbase app is left. )
Edited by: 997328 on Jun 7, 2013 2:05 PM
Edited by: 997328 on Jun 7, 2013 2:24 PM

997328 wrote:
Thanks for pointing out the "error" in the subject i changed it.
Yes I have tried it.
The thing is, in the output file there is no way to determine which member of the HSP_Rate dimension belongs where in the data. That being said there is no way the load rule will validate either. Here see below.
HSP_InputValue     HSP_InputCurrency     HSP_Rate_USD     HSP_Rate_RMB     HSP_Rate_CNY     HSP_Rate_PLN     HSP_Rate_EUR     HSP_Rate_GBP     HSP_Rate_MXP     HSP_Rate_INR     HSP_Rate_THB     HSP_Rates
Jul     FY13     BA     Working     Local     Stat_Center     230     xxx - CC10     4210C     100
Jul     FY13     BA     Working     Local     Stat_Center     230     xxx- CC10     4210M     1000
Jul     FY13     BA     Working     Local     Stat_Center     230     xxx- CC10     4250M     -100
Jul     FY13     BA     Working     Local     Stat_Center     230     xxx- CC10     4299M     -132
Jul     FY13     BA     Working     Local     Stat_Center     230     xxx- CC10     4501M     0
Thats just first few lines of the data. But how can it tell. You see what I mean??
Edited by: 997328 on Jun 7, 2013 2:23 PMActually there is a way to know what value is what member. The line
HSP_InputValue     HSP_InputCurrency     HSP_Rate_USD     HSP_Rate_RMB     HSP_Rate_CNY     HSP_Rate_PLN     HSP_Rate_EUR     HSP_Rate_GBP     HSP_Rate_MXP     HSP_Rate_INR     HSP_Rate_THB     HSP_RatesIs a listing of the data values in order. It would start with the first numeric column after the members. from the look of the sample data it looks like these are all hsp_inputvalue. scroll through the file or import it to excel and look at the columns to see if any of the other columns have numeric values.
To be on the safe side, create a dummy file that has the dimension names like
Period Years ???? Version Currency entity ????? ????? ?????? then all of the hsp values and use that to build your load rule, that way if there is a row 20000 rows down that has more than hsp_inputvalue, you won't get an error when trying to load the file

Essbase RPD question - 2 columns

We build an Essbase BSO cube with historical data. Linked it to OBI (nothing special in the RPD done).
Then we put a filter onto a column: FILTER("FinHist"."FinHist - measure" USING ("Years#1"."Years - Default" = '2006'))
This works.
BUT when we add a second column with e.g. FILTER("FinHist"."FinHist - measure" USING ("Years#1"."Years - Default" = '2005'))
then we get no results at all - even though both FILTERS alone give a result.
(we are on OBIEE 11.1.1.6.7 and Essbase 11.1.2.1)

In Oracle BI you'd typically fix this problem by creating a join between the two datasources using Process Key as the key. Can you do something similar with GRC?

Essbase studio question

Hi, I am using Essbase studio 11.1.2,
I am finding that when i make changes to a hierarchy I can refresh the cube schema and my hierarchy changes are brought into the cube schema.
When i refresh the essbase model the hierarchy changes are not brought into the essbase model, and it is necessary to regenerate the model.
is this a known limitation of the studio or am i doing something wrong with the tool.
The change i made to the hierarchy was to add attributes.

From the new features guide found at
http://download.oracle.com/docs/cd/E17236_01/epm.1112/est_new_features/est_new_features.html
, recreating or rebuilding an Essbase model is not required when you perform these operations:
Change the binding, filter, sort order, or alias set bindings of a dimension element
Change the binding, range, or alias set bindings of a derived text measure
Change the value binding or ID binding of a text list
Change an overridden data load binding in a cube schema
Note that recreating or rebuilding an Essbase model is required when you perform the following operations:
Reorder, add, or remove members in a hierarchy or measure hierarchy
Add or remove hierarchies from a cube schema
Add or remove any loose measures in a cube schema
Change the measure hierarchy in a cube schema
Override the default data load bindings in a cube schema
So it sounds like it is working as designed, you need to rebuild the Essbase model

Essbase Restore question

We had 2 types of backups: file systems and application archives taken via EAS and/or MaxL. Due to some miscommunication during an acquisition, we lost our application archives. We do however have our filesystem backups intact. Our users wish to restore a couple of essbase applications including all the objects (outline, calc scripts, rules files etc) - is this even possible using the filesystem backup? How would one go about achieving it?
Thanks.

If you have all the files for the application (including index and data) then you should be able to just stop the application and restore all the files from backup.
Cheers
John
http://john-goodwin.blogspot.com/

Essbase Cube question

Users want to know the last time when the cube was run. Do you know of any other easier way other than checking the Essbase and EIS logs. Is there any MAXL or essbase command that can be used for this purpose? Thanksande

You can go to 'Database'->'Information'-> 'Modifications'in the database menu and see when the cube has been loaded last time, calculated and any outline changes.Is that what you are looking for or you want to know when a user has loged in last time?

Essbase Adapter Question

Hi All,
A small problem (I think it is): in the Workbench I only have one Essbase Adapter, however in the Web client I can see 2. One of them is showing as Global.
Any help appreciated.
Thanks
Mark

Sorry, I should have been more precise!
I am creating a new FDM application that involves integration from an Oracle ERP. In the Workbench I have added 3 adapters: ES11x-G4-E, PULL-G4-B and ERPI-FIN-A. When I go to update the Control Table in the adapter drop-down I can see 2 Essbase adapters:
[Global] (ES11x-G4-E) and ES11x-G4-E
Interestingly, when I select the non Global adapter the selection reverts to the Global option.
Many Thanks
Mark
PS v11.1.1.3

Essbase API question?

Did anybody know which patch has correct API "EssListDbFile()" in Essbase 6.1 or later.Thanks!

Hi,
There isn't a setting in the JAPI to not use SciNote by default. What you're doing is good.
The only way I know of to replace #Missing with 0 is to parse and replace. The object doesn't have a property to set for that AFAIK. You probably have a Cartesian parse already, just do it there.
public static boolean isNumeric(String str){
    for (char c : str.toCharArray()){
        if (!Character.isDigit(c)) return false;
    return true;
}Regards,
Robb Salzmann

Essbase architecture question

I was setting up a test environment for Essbase 11.1.2.2 and was wondering if anyone can think of any negatives to having everything (SQL Server 2008 repository, Foundation Services, Provider Services, Weblogic App Server, Essbase) on just one physical server. The server itself is a reasonably powerful (Xeon E5345 @ 2.33GHz, 2 quad-core processors, 32GB RAM). Since we don't use Planning or any other product, I was wondering if we can get away with this architecture in production as well without too many performance issues? Anything we need to be wary of or keep in mind? Thanks in advance.

Thank you for taking the time to respond. We have 3 monster BSO cubes (over 500GB each) and 3 other smaller cubes which fit in memory. And 50 users (usually never more than 10-15 concurrent users). The large BSO cubes seem to be good candidates to move to ASO, hence the upgrade to 11.1.2.2.

Essbase Newbie Question

Hi all,
I've downloaded/installed Essbase 9.3.1 (server, client, admin/shared/integrative/provider srvcs). Was able to get all configured and running properly. Created a very simple BS app/cube with 5 dimensions (Year/Accounts/Market/Product/Scenario). Was able to load some data via rules file without any major hiccups. WHEW! (Roskey's book has been a big help).
My problem is that the data I loaded is not matching the figures in the .txt file. Under my Senario Demension, I have an Actual member that represents 'Revenue'. When previewing the data, or accessing via excel, the amounts are much smaller, and appear to be unique counts instead of a sum of the revenue figures. I assume this has to be some type of setting/config of the Outline as the only calc I've run is the default.
Would apprieciate any advise someone can offer a newbie. Thanks.

one of the biggest downsides to add to existing values is that if the load fails in the middle or you need to reload data, you have to figure out how to get to the state before loading (restoring backup, loading data to a holding area that gets cleared first, etc). That said, sometimes it's necessary to to use add to existing values, especially if you had an extract that has a column that is not being used.

Essbase Report question...

Ok, I have something that I think should be simple to do, but am driving myself nuts not being able to figure it out. How do you write a report script so that the first line in your output file has the dimension names as the column headers? I can get the report to dump the data in the format that I want, but I can't get it to label the dimensions defined in my <ROW, it only labels those defined in the <COL. If I could get it to label all columns, then I could use the "Record containing data load field names" option in the load rule to load the exported data back, but since it only labels some columns, I can not. Basically, I'm getting... WK1 WK2 WK3 ...ABC 2002 XYZ "Inv Qty" 10 20 30 ...I want...Product Year Customer Measure WK1 WK2 WK3... ABC 2002 XYZ "Inv Qty" 10 20 30 ...What is the syntax to get it to label the dimensions that make up my row? Thanks in advance for the help!

What is the syntax of your report script?

Hyperion Essbase upgradation question

Hi viewers my name is prasad.. i need one help right now we are in the doing the up gradation of essbase11.1.1 to 11.1.3 but we don't have an experience in the previous. can anyone guide me that what is the precautions we need to take for that and how to take the back up for that 11.1.1 while we are doing the up gradation .

Nothing is like rocket science yeah u need to test in ur dev server then i think u would be able to do ,...give a effort first
first back up ur server and apply ..what i mean is start installer and there should be step or option where u want to upgrade ...hmm dont remember exact terminology

Essbase Clac Question

I have the following structure of a BSO Cube:
Accounts
Time
Dense1
Sparse1
Sparse2
Sparse3
Sparse4
I need to roll up the cube for One level 0 member of the accounts dimension, a given month and a Given Dense1 Member (thse are all level 0).
What would be the best way to go about this ? Given that all dense dim is dynamically calculated.
Thanks,
BB
Edited by: user9519141 on Aug 9, 2010 11:38 AM

As mentioned by Glenn, there wont be much of impact of using IF or FIX. its better to use FIX in this case.
Have you considered removing non aggregating sparse dimension from aggregation ?
there are many factors that affect aggregation performance
like outline order, correct sparse/dense settings, efficient usage of storage properties, calc script optimization, cache settings
you might look in to reordering outline in the below order
largest dense to smallest dense
smallest sparse to largest sparse
non aggregating sparse dim
you can further optimize the calc scritp using set calctaskdims 2 or 3 option..
we cannot further suggest without knowing your outline structure, block size, cube size, cache settings etc...
- Krish

Essbase Fragmentation Question - Average Fragmentation Quotient

Similar Messages

Maybe you are looking for