http://onlineAppsDBA.com

can u plz clerafiy the diffrence between Shared Application Tier and Shared Appl_top

fadi

 

Fadi,
as mentioned above, Application Tier Sharing includes sharing APPL_TOP, COMN_TOP, ORA_TOP i.e. whole application tier.

where as shared APPL_TOP means just APPL_TOP file system sharing .

Earlier it was just Shared APPL_TOP which later extended to sharing entire Application Tier.

Atul

 

thank you atul for reply but still i am not able to understand which is better for me should i go with sharde appl top or sharde file system.

thanks
fadi

 

Fadi,
Looking at your comments, It seems you are not clear with Shared Application Tier funda (No offence to you pls)

shared appl top or shared file system are same thing , you achieve shared appl_top by sharing file system.

If you are talking about shared Appl_Top & shared Application Tier , go for shared Application Tier as this is now a days followed.

Atul

 

sorry atul,
i was expecting a clerificaiton on both of (shared Appl_Top & shared Application Tier what r the diffreances between them) so i can chose what to follow its not just like go for this or go for that.

keep up the good work
fadi

 

Fadi,
As I mentioned shared APPL_TOP is sharing only APPL_TOP where as in shared Application Tier you share all three application tier tops i.e. APP_TOP, COMN_TOP, ORA_TOP . Earlier apps dba's were not configuring shared application tier as it was not supported . Now shared application tier is the only option left , you don't see shared APPL_TOP alone. Though people still call shared application tier as shared APPL_TOP(totally out of ignorance and becoz of similarity in concept).

Atul

 

atul,
thanks
fadi

 

Hi, I am unable to log on via Self service /oa_servlets/Appslogon page after sharing the appl top. I keep getting your session is no longer valid. Have you got a working example of migrating an existing 3 tier system to shared appl top.

 

Ravi,
Which note you followed ?
I suppose some server ID or cookie is not set properly. Can you directly login to forms via dev60cgi/f60cgi ?

Check metalink note
233428.1 Sharing the Application Tier File System in Oracle Applications 11i as mentioned in

Good Metalink Notes

 

Hi Atul,

Thanks for ur tremendous help
can u plz list the differences between $ORACLE_SID and $TWO_TASK

Thanks,
sudha

 

Sudha,
TWO_TASK mocks your connection descriptor to connect to database More on this at
Whats is TWO_TASK

SID is to identify database Identifier i.e. If SID is set it will try to connect to database locally .

If both SID & TWO_TASK are set , prefence will go to TWO_TASK & it will connect as network using TNS Alais defines as TWO_TASK in tnsnames.ora .

Atul

 

Can you use ocfs2 for shared file system - I heard gfs stunk

 

Hi Atul,

What kind of file system is supported for shared APPL_TOP? If we use nfs, does crate a single point failure?

Thanks,

 

Nice Questions,

What kind of file system is supported for shared APPL_TOP?

Any file system which you can mount from to all nodes in shared appl_top is fine.

If we use nfs, does crate a single point failure?
Yes, thats true you are intoroducing single point of failure , here is alternative .

Lets assume you have node1, node2, node3, and node4 .

Create a filesystem share between node1 & node2 and create a snapclone/copy of that file system & mount it across node3 & node4 . After patching represent filesystem from node1,node2 to second file system on node3 & node4 .

That way you can avoid from single point of failure.

Atul

 

baddabingtim,
Yes you can use OCFS2 as well. For others who don't know about OCFS2 , its Oracle Cluster file system .

This can also be used to configure Standalone Database RAC but its not certified for RAC with Apps.
For RAC with APPS ASM (Automatic Storage Management ) is recommended & certified.

Atul

 

Hi Tech whiz,
Which version of apps you are using ?

Did you follow any metalink note ?
If not check if you completed task mentioned in metalink note

233428.1 Sharing the Application Tier File System in Oracle Applications 11i

If even after completing steps in above note then check whats error message in $OAD_TOP/conf/$CONTEXT_NAME/iAS/Apache/Apache/logs

please note your Apache log files changed now from $IAS_ORACLE_HOME/Apache...

to

$OAD_TOP/conf .....or any conf directory you created during shared Appl_top

Now you can raise all these issues in forum dedicated to Apps DBA's at

http://teachmeoracle.com/forum

 

Hi Atul,

Can you provide me the steps of installation of Oracle 10g RAC with full structure and in explanation?

 

Rahul,
Check this beautifully written doc on RAC by Jeffrey Hunter

http://www.oracle.com/technology/pub/articles/hunter_rac10gr2_iscsi.html

 

Hi Atul,
I want to clarify a dowt reg whether Refresh & Clone are same are not if not what is diffe b/w them.
Vikram

 

Yes, they meant to be same things , creating replica of source instance

 

How can we monitor the patch progress
in oracle 11i?

Very Good Blog for Apps DBA's .
Thankyou atul.

 

How can we monitor the patch progress
in oracle 11i?

Very Good Blog for Apps DBA's .
Thankyou atul.

 

How to monitor the patch progress in oracle11i?

 

Hi Atul,

What happens internally in Shared application file system...like how the forms,web services communicate?

Thanks

 

Hi Atul

In Shared Application tier configuration, we have been seeing apache crash after generating a core file.

Our server nodes are Solaris Zones and we have 3 web nodes for production database load balances using F5.

 

have you ever implemented staged appl_top?if yes, what are the issues faced while implementation or while applying patches etc. just curious as i have to implement one soon.

 

A comment regarding NFS: version 4 supports replication. You can have a setup whereby a client looks for node1 and then node2 if node1 is unavailable. However, you need a method to sync the two nodes :(

The company I work for is creating 2 separate APPL_TOPs in separate network zones. One APPL_TOP will be accessed by outside vendors and reside in a DMZ. The other APPL_TOP will be sitting within the normal production environment along with the database.

The vendor APPL_Top will be a subset of prodcution's APPL_TOP. This setup requires twice the maintenance because any changes to one has to be made the other.

Is there any way to have a single APPL_TOP replicated across separate networks? Sort of like a cluster or mirror?

 

Atul,
This is in reference to the shared apps tier note, this is my current setup,
Node 1: Db + Admin + CP + Reports
Node 2: Forms + Web + Disco

My question is when applying apps patches, do they need to be aplied on both tiers ? or since the file system is shared do we need to run auto-config ? Or just applying the apps patches on the admin tier would suffice ?

TiA,
-raghav

 

Atul,
This is in reference to the shared apps tier note, this is my current setup,
Node 1: Db + Admin + CP + Reports
Node 2: Forms + Web + Disco

My question is when applying apps patches, do they need to be aplied on both tiers ? or since the file system is shared do we need to run auto-config ? Or just applying the apps patches on the admin tier would suffice ?

TiA,
-raghav

 

hai
what is merging of APPL_TOP?

 

Can anybody tell me the command or steps to find my Appl_top is shared or not

 

Hi Raghav,
Yes, you need to apply a patch twice, i.e on both the nodes separetely, although it is in shared merely running it on admin tier and running of autoconfig wil not help.

Regards,
Phani

 

How can i find appl top is shared or not ?

 

I had done the shared appplication installation on vm's. Setup is like this on 1) vm - db is installed , 2) vm - application tier -- enabled all the services 3) vm is shared of second vm appl tier -- all the services are enabled . I had doubts

1) If I access the url of 2-VM it redirects to the 3-vm (shared) on . Can we have two differents url's opening at a time pointing to its own VM? ex. http://r12apnd.test.com:8003 --2 VM & http://r12apnd2.test.com:8003 -- 3VM. IF I access r12apndvm url it takes to r12apnd2 url.

2) Configured the PCP, how to check whether the configuration is working ? If I check the FNDLIBR on both v's only 1 VM is showing the FNDLIBR process (ps -ef|grep FNDLIBR ) other VM has no FNDLIBR process. Is it possible to have FNDLIBR processes on both the nodes (Vm).