This post I am going to review Day2 of Oracle Apps Upgrade Training where we cover Upgrade from both 11i & R12.0/12.1 and also cover Oracle Database Upgrade and Architecture of new R12.2 (Including Dual File System and New Patching Tool ADOP)
For Feedback of Day1 12.2 Upgrade Day 1 Training check Oracle Applications R12.2 Upgrade training where we also shared the questions asked by trainees in Day 1 session. Hope you learned something new by reading my last post on Oracle Apps DBA (E-Business) R12.2.5 Upgrade Training : Day 1 Review and hope this helped you in clearing some of the concepts.
Today in this blog we are covering snippets of the Day 2 session of the R12.2 Upgrade Training.
The Day 2 of the R12.2 Upgrade training started with answering questions from revious weeks session and then deep dived into the left out topics of previous weeks session before continuing with the Day 2 session.
Today we are going to talk about key points that Amit covered in Day 2 session of Oracle R12.2 Upgrade Training. Day 2 of the workshop started with covering the ADOP cycle (ADOP is new patch tool in 12.2 , replacing ADPATCH in 12.1, 12.0, 11i) and key comparisons between R12.1 and R12.2 and most importantly the features of R12.2 from an Apps DBA perspective. Once we are done with the left over topics, we immediately deep dived into the actual Day 2 session which covers the Upgrade Flow and Deep Dive of stesp involved with it
And here is one of the PPT from the session
Covering 6 topics on the R12.2 upgrade from Day 2 PPT
- Upgrade Flow
- Upgrade Paths
- Upgrade Sizing
- Database Upgrade Overview
- Preparation Tasks
- Upgrade Driver Run
We covered the 9 step upgrade process which the DBAs needs to follow for an R12.2 upgrade and also the various upgrade paths that is available for the various versions of E-business Suite.
We were supposed to cover the upgrade path up to running the upgrade driver run but because of lots of questions (again we encourage to ask lot of questions in live sessions), we only managed to scratch the surface on the upgrade front and add one more sessions to complete the remaining topics
Here are just some of the questions that we discussed on Day2 and worth sharing
Q.1) In Multinode setup can u tell us which is shared in shared appl_top ?
A.1) In case of R12.2, everything including the instance top is shared, Thanks to Oracle’s advancements in Fusion Middleware technology, you no longer need a local filesystem to hold your Apache configuration and log files, which means that everything is truly shared across all the nodes.
Q.2) Can you briefly explain about FS_NE – FS_NE stands for Non Editioned Filesystem ?
A.2) Files which requires persistence and cannot be swapped between the run and patch filesystem (There are two file system in EBS 12.2 i.e. RUN & PATCH and this new File System We cover in Day1 of Training) stays in the Non-Editioned filesystem, for example, concurrent manager log and output files, patch log files, adop utility and the patches itself
Q.3) Whenever we change APPS password should we change the credentials in datasource as well ?
A.3) yes, we have to do that.
Q.4) Do we need to change this password in password in patch edition data source as well ?
A.4) No, ADOP will take care of that.
Q.5) Code migration should be deployed on both FS1/FS2?
A.5) No, we need to deploy only on RUN filesystem, again ADOP will take care of synchronising patch filesystem when called.
Q.6) Does adRegisterWLListener.pl always runs in the background ?
A.6) Yes, it always runs in the backgrounds, start up when the admin server is brought up and shuts down when the admin server is brought down
Q.7) What is difference between adRegisterWLListener.pl and adSyncContext.pl ?
A.7) adRegisterWLListener.pl will automatically synchronize only very few changes from FMW to context file whilst adSyncContext.pl does complete synchronization, whenever any changes are done to FMW Home, it is recommended that adSyncContext.pl is executed to push the changes to context file.
Q.8) How would you compromise on cpu/mem on node, when we apply a huge patch in patch filesystem with parallelism ?
A.8) When you are applying a patch using adop cycle, you should always remember that end users are still accessing the environment and also the fact that you do not have any restrictions to complete the patch ASAP, the DBAs should go with an optimal number of workers depending on the environment, so that end users are not impacted
Q.9) Cross edition triggers, should developers create it for custom code ?
A.9) No, as long as the table is created as per Oracle R12.2 standards and properly setup, Oracle will take care of creating the cross edition triggers
Q.10) If we query the fnd_nodes synonym in APPS schema will it fetch results from fnd_nodes# edition view or fnd_node table ?
A.10) It will fetch results from the fnd_nodes# edition view, direct table access is not recommended anymore in R12.2 scheme of things.
Q.11) What is difference between finalize_mode=full & finalize_mode=quick ?
A.11) Full finalize mode will gather statistics for internal tables which will speed up adop cutover but only required if we are applying a huge patch or lots of patches, otherwise doesn’t matter.
Q.12) How to restart if patch failed during apply ?
A.12) you have to restart using adop utility with keyword “restart=yes”
Q.13) What is the difference between fs_clone and prepare phase ?
A.13) Prepare phase internally calls fs_clone if it is not able to synchronize the patch filesystem incrementally. fs_clone is an one off process and is not considered as part of an ADOP cycle whilst prepare phase is the first step in an ADOP cycle.
Q.14) Old editon will be obsolete after ADOP Cleanup phase ?
A.14) Any old edition which is discarded with an aborted ADOP cycle will be cleaned up, any editions which is parent of an existing child edition, will remain on the system and will not get purged. However obsolete coded objects will be removed those editions.
Q.15) In case of multi node do we need to run adop prepare,apply,finalize etc on all nodes or only one node ?
A.15) We have to do all the phases only from master node, it will take care of firing the relevant commands on the slave nodes
Q.16) What is the difference between fs_clone and fs_clone with parameter “force=yes” ?
A.16) A fs_clone session is called with “force=yes” parameter if the previous fs_clone session failed for some reason and you want the new fs_clone session to start from scratch. If you want it to continue from where it left earlier, then you should not pass the force parameter (default method)
Q.17) we have a custom database schema, after moving custom reports from old to new environment, what are extra steps to complete custom_top enablement?
A.17) The custom schema must be registerd using adsplice utility and we have to ensure that Oracle development standards are followed for all your custom objects.
Q.18) How to handle custom exceptions in Online Patching Compliance Report ?
A.18 ) Most of the exceptions might go once you register your custom schema using adsplice utility however some of your customization would have to rewritten as per R12.2 Online Patching standards, this is manual work and would have to be done by the developers.
I am interested How can I learn more and Quickly ?
If you want to learn Oracle EBS Upgrade R12.2 (Financials, HRMS, SCM, Technical) Upgrade to R12.2 at a faster pace with tons of features like Dedicated Machine to Practice, Live Sessions, Support, Retaking Sessions for next 1 Year and much more. Then you can also check out our paid workshop by visiting https://k21academy.com/oracle-ebs-upgrade or on blog here or
Did you download FREE 7 Docs every Oracle Apps DBA must read before Upgrading to Oracle R12.2 ?
Having a question related to your Oracle E-Business Suite Upgrade to 12.2 or Training for Oracle ACE & Author, Atul Kumar then post under comments below for instant reply.