Oracle Enterprise Manager Ops Center 12c Update 1 - Additional Information and Best Practices
By Anand Akela on Jul 23, 2012
Oracle Enterprise Manager Ops Center 12c Update 1 was released earlier this month. Eran Steiner , Technical Architect, Oracle Enterprise Manager, adds some additional information and best practices about upgrading to Ops Center 12c Update 1 in this blog.
Eran hosted a call to provide an overview of Oracle Enterprise Manager Ops Center 12c Update 1 and answer
any questions.The recording of this call is available here and the presentation can be downloaded here.
- Upgrade, Update and Install - Update 1 is a multipurpose bundle: it allows you to update your 12c GA version to Update 1, upgrade your 11g U3 environment to 12c, and it also allows you to do a fresh install straight into Update 1.
- Oracle VM SPARC Server Pool HA - with Update 1, if you configured a server pool to support it, Ops Center can automatically recover LDOM guests (also known as OVM for SPARC) to a surviving Control Domains in case of a failure. In other words, if the machine hosting the guests has a failure, Ops Center will automatically recover the LDOMs to a working machine. In addition, a new option was added to determine how to treat the failed Control domain OS:
In most cases, you will want to enable this to avoid a scenario called "split brain" - where you have the same guest running on two different servers.
- Oracle VM SPARC IaaS (Virtual Datacenters/clouds) - previously you could create Virtual Datacenters using Zones and Oracle VM (OVM) guests. With Update 1 you can now create Virtual Datacenters (clouds) also using LDOMs. Here is a screenshot where the 3 types coexist under one Enterprise Controller. Of course, you can assign different users to each of these clouds:
- WANBoot Improvements with OBP Handling Enhancements - Update 1 fixes a number of issues around WANBoot and now allows you to also use WANBoot when provisioning LDOM guests
- SPARC SuperCluster Support - a few enhancements to SPARC SuperCluster features
- Stability fixes - many bug fixes and small enhancements across the board
Planning the upgrade
Before starting the upgrade, make sure to review the Upgrade documentation:
As well as the release notes:
Where to download Update 1
You can download the Update either in the User interface or, if you want to perform a manual installation/upgrade, through eDelivery.
In the User Interface, go to Administration -> Ops Center Downloads -> Choose the 2076 bundle, that's Update 1.
After the download is complete, you can find the bundles here:
Enterprise Controller upgrade – /var/opt/sun/xvm/images/productUpdate
Proxy Controller upgrades – /var/opt/sun/xvm/images/productUpdate
Agent Controller upgrades – /var/opt/sun/xvm/images/agent
Note: if you choose to upgrade through the user interface, you do not need to access these folders. These are only helpful if you are planning on performing a manual upgrade and need to access the upgrade ZIP files.
To download through eDelivery, go to https://edelivery.oracle.com/ , login and then under "Select a Product Pack" choose "Oracle Enterprise Manager" and the relevant platform and click "Go".
From the result list, find "Oracle Enterprise Manager Ops Center 12c Release 1 Media Pack", or Release "188.8.131.52.0" and click it. You will see the bundles next.
Tips for 11g installations
There are a couple of key differences between the 11g and 12c products that you should be aware of prior to upgrading:
- New requirements: make sure you verify the new requirements by running the "OCDoctor --troubleshoot" - not only will it alert you about existing issues with your installation, it will also tell you if anything is missing for the upgrade. More on the Doctor below.
- Additional information about requirements can be found here:
- New database: 12c is using Oracle DB which is different than the PostgreSQL used in 11g. You can have the Oracle DB installed on the same server as the Enterprise controller (called "local" database) or you can use a remote Oracle database. You can choose which way to go based on your needs. If you upgrade through the HTML user interface, you only have the option to use a local database. If you upgrade manually through the shell, you can either upgrade to a local database or a remote database. There is a method that allows you to move from a local database to a remote database, if you choose to do so later on.
- Before the upgrade, make sure you download the required Oracle DB installation files. Download instructions are provided in the Upgrade documentation link above.
- Before you start, we strongly recommend you perform a backup. Not only you will have a copy of your environment saved, it also restarts the services in the process and will have the EC services fresh and ready to go. 11g Backup documentation is here:
Step by step highlights
Here are some tips before you start:
- We strongly recommend you perform a backup. The Upgrade process is designed to perform a backup as well, but it never hurts to have one outside the upgrade process and to make sure it works well. 12c Backup documentation is here:
- OCDoctor - The OCDoctor utility is updated very often with all the known issues we hear from support and folks like you. It can point out potential issues, possible workarounds or even automatic fixes. We cannot stress enough how important it is to run the OCDoctor before and after the upgrade. It can really make your life a lot easier.
- Before the upgrade, run the following on your Enterprise Controller.
# /var/opt/sun/xvm/OCDoctor/OCDoctor.sh --update
This will get the latest version.
# /var/opt/sun/xvm/OCDoctor/OCDoctor.sh --troubleshoot
The above will check for current issues and recommendations. If you are on 11g, it will also check for the requirements of the 12c.
- After the upgrade, run the following again:
# /var/opt/sun/xvm/OCDoctor/OCDoctor.sh --troubleshoot
The above will check for any known post-upgrade issues.
- If this is a new install, you can get the latest Doctor here:
Then run the following on the designated machine to check for the prerequisites:
# unzip OCDoctor-latest.zip
# cd OCDoctor
# ./OCDoctor.sh --ec-prereq
- Upgrade the Enterprise Controller first. This will also upgrade the collocated proxy and agent, if you have them installed on the same machine as the Enterprise Controller.
- Next, proceed and upgrade the remote proxies. If you have proxies on different architecture than the Enterprise Controller, you may have to download the Proxy upgrade bundle for that architecture as well (either through the UI or get it from eDelivery).
- Finally, it is strongly recommended that you upgrade all the agents as soon as possible.
- Important: In 12c, the upgrade agents screen has moved! It is no longer under the Administration tab. Instead, under "Assets", change "All Assets" to "Operating Systems", choose the right OS group (for example, Solaris), click the "Summary Tab" and then "Agent Controllers"
Some additional tips
- If you are upgrading from the User interface, once you start the upgrade you will only see an Upgrade splash screen with no status updates. Since the upgrade can easily take 30-60 minutes, you can view the progress of the upgrade through the following log:
# tail -f
After the upgrade is done, the install log can be found here:
- If Update 1 does not show up in the HTML user interface, try restarting the following services:
# /etc/init.d/uce_scheduler stop
# /etc/init.d/uce_server stop
# /etc/init.d/uce_server start
# /etc/init.d/uce_scheduler start
# svcadm disable uce-scheduler
# svcadm disable uce-server
- wait a few seconds to allow services to shut down -
# svcadm enable uce-server
# svcadm enable uce-scheduler
Then wait for a few minutes for the Knowledgebase to refresh and try again.
- Disconnected customers (offline customers) can download the bundles from eDelivery and upgrade manually - follow the same steps for command line upgrade.
Good luck with the upgrade and with new installations!
Contributed by Eran Steiner , Technical Architect, Oracle Enterprise Manager