Cross-posting what I wrote on my LinkedIn.
"Glad to see so many questions quietly coming in via private chat about Oracle EPM 11.2. I encourage anyone considering an on-premises upgrade in 2021 to spin up a VM now and make sure you can successfully get through RCU and the EPM Configurator. Don't wait until your project is scheduled to start in 2021.... act now while things are quiet and you still have time to troubleshoot. This will go a long way to have a successful project in 2021!
Please do NOT send any diagrams or logfiles to me unless we have non-disclosure and SOW paperwork in place."
One additional point I made in private chat concerning FDMEE problems.
If you get errors about missing SNP_* tables when starting up FDMEE, the installation media is the likely culprit. Open a SEV-2 SR with Oracle so they know about the issue.
While you wait from a response from Oracle, you could try this approach:
Create a new VM that is specifically for FDMEE. Download EPM 11.2.2.0 and do a fresh FDMEE install on that VM. Download FDMEE 11.2.2.0 as soon as you read this post, as I question the 11.2.3.0 base installation media for FDMEE/ODI and it will eventually get removed from eDelivery.
You do not need to configure or RCU because this VM will be thrown away once you get through this exercise. Then throw the latest EPM 11.2.x (11.2.3.0 as of this writing) on top of it and do a Maintenance in-place update.
Finally, zip up Oracle\Middleware\odi\ on that VM and copy it over to the troublesome FDMEE machines. Re-run the configurator and pick "Configure Database" for FDMEE. Then check your database and verify the many dozen ODI tables are now present.
Your results may vary. Back up the odi filesystem on the FDMEE machine(s) before you replace it. Keep in mind the above is a suggestion and might not have Oracle's blessing at this time.
Hi Dave,
ReplyDeleteHave you been able to deploy EPM 11.2.3 in a distributed environment without issues and/or even 11.2.3 fresh install in a standalone mode on LINUX?
Regards,
Olivier BENNARDO.
Hi Olivier and welcome,
DeleteYes, I have done a fresh install of EPM 11.2.3.0 in Red Hat Enterprise Linux (RHEL) 7. I've also done distributed 11.2 installs, but not yet in RHEL7. The RHEL7 install I performed was a single node with the relational database running on a different node.
In RHEL7, 11.2.3.0 all worked fine except for FDMEE/ODI. Time permitting, I'll want to go back and start from scratch with a distributed RHEL7 install to see if FDMEE/ODI is still problematic, or if Oracle fixed the installation media on eDelivery.
If you're just using Essbase/Planning, you'll be just fine.
As of this writing, HFM is still unavailable for RHEL7 so you'll need a Windows server (2016 or 2019) in the mix. The last update I received is HFM for RHEL7 is on Oracle's development roadmap.
Best of luck,
Dave
Hi Dave,
ReplyDeleteAll seems to work fine, but we are struggling with EAS login due to Weblogic OPSS credentials not created for EAS Server on our install. Have you faced this issue?
If you're referring to the EAS web launcher rather than the locally installed EAS Console thick client, there's something you need to read that's buried within the 11.2 readme. For EAS in particular, there's a WebLogic policy that needs to be edited. Grab a cup of coffee and then crawl through the readme. You'll spot what I'm talking about. I hope this solves your issue!
DeleteHi Dave,
DeleteFor EAS Console in JNLP, we managed to make it working when DOING distributed environment setup! When you are doing a full EPM standalone deployment, this is where you have to update weblogic.properties file to make it unlimited.
For now we have been successfull in the following setups:
- EPM 11.2.3 and Oracle DB on Linux 7, in standalone AND distributed environment.
- EPM 11.2.3 and SQL Server 2016 SP2/2017/2019 on Linux 7, in standalone environment.
Findings:
- You must install one RCU setup per WebLogic deployment machine (e.g. when you have to install Managed Servers)
- You must make sure RCU.properties is modified and publishing the good tags for JDBC for Oracle DB and SQL Server
- Same applies to reg.properties
- When it is freezing, it is mainly due to OPSS schema init stopping because of JDBC or password issue ==> When you made a mistake on JDBC string passwords in RCU.properties OR in reg.properties...
We are now validating that it is running as it should for the following: EPM 11.2.3 and SQL Server 2016 SP2/2017/2019 on Linux in distributed environment.
P.S:
Distributed environment = 1 Managed Server per machine, e.g., one Essbase Admin Services machine, one FR machine, one Planning machine, one APS machine, etc...
Now I confirm that it is also working for EPM 11.2.3 and SQL Server 2016 SP2/2017/2019 on Linux in distributed environment!!
Delete