Regarding Cache in clustering

Hi,
we are using JOC Cache mechanism for java objects
In Standalone application server 10.1.3.1 it is working fine
In production we are using OC4J Enterprise edition Cluster topolgy (10.1.3.1.0)
In this case we are using multiple oc4js and those oc4js we have 2 instances and 2 jvms.
in cache wise we are using in-memory concept,
we have a senario where at run time cache object values will be changed.
i have observerved that in each jvm one cache region created.
when one cache is created then it is not reflecting in another jvm (or instance)
i want to replicate when ever one region's object modified to other jvm instances
can u please provide any solution for this problem
Regards
Mannu.

Chak,
Try to understand the concept of BI reporting here. Your reports makes use of all the tables i.e. Dim & Fact to fetch the resultset and it is the report SQL which gets cached.
If a query references multiple physical tables with different persistence times, the cache entry for the query will exist for the shortest persistence time set for any of the tables referenced in the query. So, in this case it will be just 1 Day for all the tables and complete SQLs referencing these tables.
For NQCMD command utility...Purging is very easy here...and same goes for seeding.
Like SeedTableCache.txt which will have your SQL scripts. Once this is used via NQCMD which can be scheduled using Job Manager or any other job.
nqcmd -d AnalyticsWeb -u Administrator -p Administrator -s c:\obiee\SeedTableCache.txt
Similar concept goes for any flavor of BI.
Here John has given some good examples on Cache Mgt.
http://obiee101.blogspot.com/2008/03/obiee-manage-cache-part-1.html

Similar Messages

Global cache in clustered environment

We have a clustered external facing portal application with four servers and each server has two nodes. We want use to IcacheService to store some objects in a cache. But this cache is specific to each JVM and cannot be shared across the servers.
Is there any global cache mechanism in SAP through which we can share the objects across all the servers in a clustered environment?
Thanks in adavace.
Ram

Hi Rambhupal,
Would you share any solution/information for the same as we have a similar requirement.
Regards,
Melwyn

Querie regarding cache sizes

For optimizing the calculation script i have set in my cube & compression type as RLE (prior calculation script was running with time span of 6 minutes now it takes 2 minutes ,datafile exported using dataexport is same )
The maximum index cache you have set as 4097152 kb ( i.e.3.9 gb) Is it ok to set the index cache so high even though my index file size is less than 1 GB.
1)How do i conclude the maximum value for datacache is 36000000 kb. What are the factors i need to take for consideration.
2)Datacache Maximum 36000000 kb
Datacache is 36000000 kb (i.e. 34.33 GB). Is it a practical approach
Regards
Shenna

Hi,
Index Cache:
The doc suggests to have- 1 MB of index cache for Buffered I/O and 10 MB of index cache for Buffered I/O !
While you can use this recommendation to start with- You're the right person to arrive at the actual figure by doing some trials relevant to your environment.
Data Cache:
Again, the doc. suggests to have- Data cache = 0.125 * the value of data file cache size
Where- Suggested Data file cache size = Combined size of all essn.pag files, if possible; otherwise as large as possible. This cache setting not used if Essbase is set to use buffered I/O.
It's prudent to do trials independently for each of the caches!
It's worth reading all the posts of the thread @ Understanding Buffered I/O and Direct I/O to understand experts' opinions !
Best of luck :)
- Natesh

Doubts regarding cache

Hi Gurus,
Please help me to understand the follwoings:-
a)Although I have created and saved ibots successfully and can see them in active ibots session,but somehow cache is not getting seeded and I can't see any records under manage cache.Please help me to understand where I am going wrong??
b)Our OBIEE presentation,BI and scheduler servers will run in a UNIX environment in production.Thus,please help me to understand what are the steps needed to be taken to make the scheduler server and ibots running in production?Currently its running fine in my local system although I am facing some issues with seeding.but what mesures need to be taken to make it run even in production unix box also?
c)What are the steps needed to be taken to provide all the users are the advantage of caching. I mean,here i have done it for user: Administrator in my local system,now what else do I need to do to make it available for all the users once its in production?
d)can we use java script like Purge.js file for purging functions in UNIX environment too?currently we are using shell script for thsi ..but if we can use Purge.js ,then I will use ibot to schedule this and purge teh cache.
Thanks a lot,
Jit

Praba wrote:
So what I don't understand is what is the significance of the Past Image here ?
A PI (past image) is a copy of the block that is known to have existed as a current (CU) version of the block at some point in time in the past - so it can be used as the start point for recovery if the node that holds the current block crashes. So PI blocks can reduce recovery time in the database.
I believe PI blocks can also be used by the holder of the PI to generate local CR copies without reference to the DRM - if the SCNs are appropriate. So PIs can also be used interconnect traffic.
Initially there was a rule that when a node wrote the current block it had to send a message to all other nodes mark their PI copies as free buffers (in effect, forget them) - although I did see a note on one occasion that Oracle might change this to convert PI blocks to CR blocks.
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
"For every expert there is an equal and opposite expert."
Arthur C. Clarke

Regarding cache size limit

hi all,
What is the maximum possible cache size set for Webcache.
Any help would be much appreciated
Regards,
Geeta

hi all,
What is the maximum possible cache size set for Webcache.
Any help would be much appreciated
Regards,
Geeta

Toplink cache and clustering and load balancer

Hi All:
I have an application deployed on AIX platform. This is developed using struts, toplink and session ejbs.
Oracle 10g is installed on two AIX machines and these are in cluster. My j2ee application is deployed on oracle 10g app server.
I will be having loadbalancer in the front of this cluster. Even now with out loadbalancer , with oracle app server clustering I am observing the request sharing between two servers.
My problem is that as I am using the toplink and it uses the cache, what will be the impact on my application on below scenerios.
1. user A connects to machine 1 and reads object using toplink. At the same time user B connects to machine 2 and also reads the same object. User A will update the same object and commit it. User B will use this object and will put this as a child of some parent object and commit the parent.
I mean to say that due to different two machines, cache will not be in sync and this may cause a major problem.
Please provide your input to this.
Thanks,
Viral

Viral,
This is what ORM and more specifically TopLink help you solve.
What you need to do is configure TopLink to use your appropriate locking mechanisms. I recommend optimistic but in some cases pessimistic locking is appropriate too.
Optimistic locking will detect if the second machine attempts to modify an object that has already changed on the database. There is also support to force the UnitOfWork to verify that the optimistic locking field(s) have not changed when using the concurrent object in another relationship.
UnitOfWork.forceUpdateToVersionField(Object, boolean)
http://download-west.oracle.com/docs/cd/B14099_04/web.1012/b10491/oracle/toplink/sessions/UnitOfWork.html#forceUpdateToVersionField(java.lang.Object,%20boolean)
This method will allow you to detect if the instance you are about to use has been modified by another transaction with or without updating its version field.
Cache sync (coordination) is intended to minimize stale cache conditions but locking is the only way to guarantee it is handled at the database level.
Doug

Need info regarding caches

Plz correct me if I am wrong
1) Index cache is buffer where the index pages are located .and Index pages consists of pointers and these pointers tell you where the data blocks are located on the disk.
2) Data Cache is buffer where essbase created to store uncompressed data blocks
3) How about .pag files ? I guess these files consists of the data block information ie data values.
4) then How about .dat files? What these files exactly contain.
5) Data file cache is buffer where essbase creates to store compressed data blocks. I read that data file cache will be created only if you are using direct i/o .. why you need data file cache as essbase is taking data directly from disk.
6) calc cache is goiing to be created to store the blocks that are needed for calculation
Please clarify me on this caches. Correct me if there are any mistakes
Thanks

1) Index cache is buffer where the index pages are located .and Index pages consists of pointers and these pointers tell you where the data blocks are located on the disk.
Yes
2) Data Cache is buffer where essbase created to store uncompressed data blocks
Yes
3) How about .pag files ? I guess these files consists of the data block information ie data values.
A file that contains data blocks. Essbase generates the data file upon data load and stores it on disk. Essbase names data files incrementally—essxxxxx.pag, where xxxxx is a number starting
with 00001.
4) then How about .dat files? What these files exactly contain.
.dat files are data files for storing data.
5) Data file cache is buffer where essbase creates to store compressed data blocks.
The size of the data file cache determines how much of the data within the data files can fit into memory at one time. The data cache size and the data block size determine how many data blocks can fit into memory at one time. Data files for a single database can span multiple volumes;multiple databases can reside on the same volume.
6) calc cache is goiing to be created to store the blocks that are needed for calculation
Yes.

Regarding Cache in obiee10g

Hi I have scenario like these Fact1,Fact2,fact3 and DIM1,DIM2,DIM3 and one more Time DiM4 table. The data in the DIM1,2&3 tables are not going to change. Every day night they load new data in the 3 Fact tables.
1.My first report is based on Fact1,DIM1,DIM2,DIM3& DIM4.
2.Second report is based on Fact2 DIM1,DIM2,DIM3& DIM4.
3.The third report is based on Fact3 and DIM4 & DIM2.
Please help me to set cache for the tables in rpd level.
Thanks,
chak

Chak,
Try to understand the concept of BI reporting here. Your reports makes use of all the tables i.e. Dim & Fact to fetch the resultset and it is the report SQL which gets cached.
If a query references multiple physical tables with different persistence times, the cache entry for the query will exist for the shortest persistence time set for any of the tables referenced in the query. So, in this case it will be just 1 Day for all the tables and complete SQLs referencing these tables.
For NQCMD command utility...Purging is very easy here...and same goes for seeding.
Like SeedTableCache.txt which will have your SQL scripts. Once this is used via NQCMD which can be scheduled using Job Manager or any other job.
nqcmd -d AnalyticsWeb -u Administrator -p Administrator -s c:\obiee\SeedTableCache.txt
Similar concept goes for any flavor of BI.
Here John has given some good examples on Cache Mgt.
http://obiee101.blogspot.com/2008/03/obiee-manage-cache-part-1.html

Regarding caching mechanism - Pls respond ASAP.

Hi All,
I am trying to apply "purging of cache" on daily basis in my application because it is on transactional database. How could i proceed. I think if i am adding a event polling table and set the frequncy to 1 day then i can achieve that. But is there is any other way to do that automatically because i don't have scope to add another table in the database.
I have change the cache expiry of all physical table as 1 day. Is that will purge the cache every day or i need the event polling table.
Advance thanks on any suggestions.
Thanks
Ashok

Thanks John.
But we have migrated the system in production environemet. Thats why we can't run the script there everyday because no one has permission to access the server.
For this we have to explicitly take access and run the script. That is not possible as far as i see the production setup.
Is there any way to automate this.
Thanks
Ashok

Need help regarding cache

what r the attributes of response.setHeader("?","?") and one more what is response.setIntHeader

You are probably looking for:
response.setDateHeader("Expires", 0);
response.setHeader("Pragma", "No-cache");
response.setHeader("Cache-Control", "no-cache");They are HTTP response headers sent by the webserver to the browser.

Regarding Cache

Hi Experts,
When we generate the report(Say for ex Col-A,Col-B,Col-C) in the answers it will show the report and weather the whole results report will store in the cache or only logical query will store in the cache.
Again when we generate the same report in the answers how the presentation server will know that this is the previous report i.e. on which basis it will compare and what are the conditions should satisfy for hitting the cache.
Thanks,

Rafi.BI wrote:
Hi Experts,
When we generate the report(Say for ex Col-A,Col-B,Col-C) in the answers it will show the report and weather the whole results report will store in the cache or only logical query will store in the cache.
Again when we generate the same report in the answers how the presentation server will know that this is the previous report i.e. on which basis it will compare and what are the conditions should satisfy for hitting the cache.
Thanks,Hi Rafi,
Read through this post which should answer most of your questions: http://obiee101.blogspot.com/2008/07/obiee-cache-management.html
Go through the comments as well.
Thanks,
-A.Y

How to create Groups and Group Leaders in Clusters.

Hi,
As we know in unicast there is one to one communication and there are groups to control the Thread
Management, How the Groups and the Group Leaders are created.
Regards,
Vardhan.

Unicast clustering uses TCP/IP sockets to pass cluster messages between members. To avoid requiring each cluster member
to have connectivity to every other cluster member, WebLogic Server uses a group leader strategy whereby the oldest member
of the group (in other words, the server that was started first) is designated the group leader. All members of the cluster
connect to the group leader so that the group leader acts as the relay point for cluster messages between members.
If the group leader goes down, the next oldest member becomes the new group leader.
As you can imagine, the group leader strategy works well for small groups but becomes less efficient as the number of members
of the group grows large. As such, WebLogic Server uses a multiple group leader strategy where it limits the number of members
in a group to 10. If the cluster is larger than 10 members, WebLogic Server splits into two or more groups, each with their own
group leader. The group leaders themselves are all interconnected to minimize the number of hops that a cluster message must
traverse to reach all cluster members.

PI Caching Problem

Hello Experts,
we are faced with a strange problem. We are using SAP PI 7.1 EHP 1 and we have developed some integration processes. Now we have transported our software component to another SAP PI 7.1 without EHP 1. Transport also was successful.
Everything worked fine. Then we did some changes in our software component and transported it again. Now we are faced with the problem that one of our four intergration processes have not been upgraded after the transport. In the sxi_cache the integration process always have the date of the first transport of our software component.
Can anybody help? All cache refreshes and a restart of the PI dont solved the problem.
We also did a copy of this integration process and activated it. The copy is never shown in the sxi_cache. It seem like the integration process doesn't exist. But he is there.
Have anybody ever had the same problem? How you solved this?
Thanks for your help in advance.

Go to http://<host>:<java port>/rwb/sub/cache_monitoring/notification_table and look into problems tab to get more details regarding cache problems on your system.
You can fix the problem by creating a new version of the object in dev and transporting it
OR
Doing a full cache refresh in SXI_CACHE

Caching works in the portal

Hi,
Can you help me regarding caching works in the portal?
The problem we are having is that when a new report is transported to BI ABAP quality, it does not show the changes through the portal. If we restart the portal system, the changes then appear.
Thanks for your help.
Cedric

Hi Martin,
Thanks for your response.
Unfortunately, that has not solved my problem. Even when I change the caching parameters of my iview, I have the same problem :
A query has been modified in the BI ABAP but I dont see these changes when I launch again the iView which calls this query.
I have logged off the session, cleared browser cache and it still has the same problem.
I think it's a portal caching issue. I have the iViews to expire the cache but it seems to be getting overrided by another settings.
If someone have an idea ?
Paramerters in my iview :
Allow client-side caching = NO
Cache Level = None
Cache Validity period = 10Min
Thanks for your help.
Cedric

Caching in FMS 2.0.4

Hi,
We have set up FMS 2.0.4 and want to find out if caching
capability is available in FMS and information regarding cache
settings. Please let me know if this information is already
available somewhere.
Any help on this topic would be helpful.
Thanks!

Try looking into the Edge configuration guide. Its a pdf that
comes with the FMS. Configure Edge and Origin server to check the
caching capabilities

Regarding Cache in clustering

Similar Messages

Maybe you are looking for