Oracle has been recognized as a Leader in the 2023 Gartner Magic Quadrant™ for Transportation Management Systems. Oracle is positioned the highest for “Ability to Execute”, and furthest overall for “Completeness of Vision”. This is the 16th time Oracle has been named a Leader in the Gartner Magic Quadrant™ for Transportation Management Systems™ report.
Oracle has been named a Leader and positioned highest for ‘Ability to Execute’ and the second position for ‘Vision of Completeness’ by Gartner in the 2023 “Magic Quadrant for Integration Platform as a Service (iPaaS), Worldwide.” The report recognized Oracle as a Leader for the sixth consecutive time.
To the report
Recently, I described that starting with Release 22.2.1, the Oracle E-Business Suite Cloud Manager now supports OCI File Storage Service (FSS) besides block storage. Furthermore, the usage of FSS is mandatory for (new) multi-node environments: First Experience with E-Business Suite Cloud Manager 22.2.1
Given the advantages of FSS described in my previous blog one of my customers decided to migrate their apps-tier to File Storage Service. Let’s see how we did this:
In preparation for the move first of all we had to create a new file system:
Creating the new file system.
This file system then has to be attached to a mount target and needs the proper export options:
NFS Export Options
Then, the mount point can be mounted in a temporary directory:
sudo mkdir /mnt/fss sudo vi /etc/fstab 10.22.9.97:/ENTW220811 /mnt/fss nfs rw,bg,hard,timeo=600,nfsvers=3 0 0
Then:
sudo mount /mnt/fss sudo chown oracle:oinstall /mnt/fss sudo yum install -y fss-parallel-tools/
After these preparatory steps, we stop the entire environment and copy the content from the previous /u01 block volume over to the new FSS mount point /mnt/fss. Performance can be greatly improved by using fss-parallel-tools and parcp for this. For volumes of considerable size, an incremental operation can also be handy; but for the 300 GB one usually used for the apps-tier doing the migration in a 1-2 hours downtime is usually not a problem:
. setenv_run.s adstpall.sh -mode=allnodes ~/stop_apex122.sh
sudo parcp --restore /u01/install /mnt/fss/
If working in a multi-node environment, then on the second apps-tier the mount through NFS to the primary apps-tier then needs to be replaced with a mount to the FSS mount target.
After the copy is finished, the old /u01/install mountpoint is retired and FSS is mounted to /u01:
sudo umount /u01 sudo vi /etc/fstab 10.22.9.97:/ENTW220811 /u01 nfs rw,bg,hard,timeo=600,nfsvers=3 0 0 sudo mount /u01
Then, start the apps-tier again:
. setenv_run.s adstrtal.sh -mode=allnodes ~/start_apex122.shh
After the downtime the old block volume should be detached through the OCI console. Furthermore, it is necessary to re-discover the environment in the E-Business Suite Cloud Manager. For this, either unregister the existing environment or (as I did in order to pick up the latest operating system image and add compatibility with the latest OCI shapes for the Cloud Manager) setup a new Cloud Manager environment. Make sure the Cloud Manager network profile is aware of the FSS Mount Target (usually, you have to create a new network profile for this).
Then, issue a re-discovery request:
Rediscovery of the moved environment
FSS is a highly durable service. However, that does not help against e.g. user errors (issuing “rm -rf /u01”). To remedy this, it is helpful to setup FSS Snapshots with e.g. an hourly snapshot (kept for a day) plus a daily snapshot (kept for a week). Unfortunately, so far this cannot be setup from the OCI console. Furthermore, for FSS it is currently not possible to perform automatic backups to OCI object storage (this is on the roadmap, though). However, using the utility fss-scheduler, such a backup policy can be easily setup.
As described in my previous post, FSS is considerably more expensive “per gigabyte” (see the oci price list) compared to block storage: 30 cents compared to just around 4.5 cents per GB and month. However due to the “sparse” nature of FSS, for an environment with production plus 3 clones for dev/test our storage usage changed from 4x400GB=1200GB with block storage (equaling roughly 54 USD/month) to 200GB + (3*15) GB = 245 GB (which equals 74 USD/month). The relative increase in costs may still seem significant, compared with the total costs for hosting 4 E-Business Suite environments with terabytes of block storage for the database; these are, however, peanuts. The ratio might even improve if in the future, Oracle will eventually use FSS clones for creating the patch file system (see, and vote for my idea over there).
By performing the above steps, you can migrate an E-Business Suite apps-tier easily to File Storage Service in a downtime of about 1 hour. By doing so, this it is e.g. possible to create very fast clones that are furthermore “sparse”.
Overview
In integration projects, it is often needed to integrate cloud applications with on-prem systems running in the clients network. In most cases, the client network is not accessible from the internet. This situation makes the task of integrating this system for a cloud integration platform, such as the Oracle Integration Cloud (OIC), almost impossible.
Fortunately, the OIC comes with a functionality named Oracle Connectivity Agent (OCA). The small Java program that can be run as a service on a linux machine closes this gap. The OCA can currently work with the following OIC adapters as outbound (invoke) connection:
As well as the following adapters as inbound (trigger) connection:
(Some) Advantages of the OCA are:
The basic pattern of how to use the OCA is the public internet pattern. In this pattern, the OAC will be installed in the customer network and communicate between the on-prem application and the OIC. The all connections will be via the public internet.
More information about the OCA can be found in the according Oracle Documentation (namely https://docs.oracle.com/en/cloud/paas/integration-cloud/agent.html) as well as in the related tutorial provided by Oracle (at https://docs.oracle.com/en/cloud/paas/integration-cloud/integrations-user/create-integration-exchange-messages-oracle-integration-and-local-host.html#GUID-F92564A0-8C6F-4C35-A5EE-B3F19571B616).
To create an OCA, you first need to login into the OIC, then navigate to „Integration“ – „Agents“ – „Create Agent Group“. The next step is to fill out the basic information for the OCA and save it.
After clicking “Create”, you should see the newly created agent in your agents list.
Next, download the agent ZIP file and copy it on the server, where the agent should run. Be aware that the agent cannot be run as a Windows service, so if you want to use the agent on a Windows server, the user must be logged in all the time. In the following, the explanation will apply to a Linux server.
After copying the files, unzip them into the folder where you want to store the data:
and modify the InstallerProfile.cfg
Check if JAVA_HOME is part of the PATH variable. If not, add it. After this, start the agent with:
If you want to run the connectivity agent as background service, you need to add
oic_USER=
oic_PASSWORD=
into the InstallerProfile.cfg and start it with
After this, the agent is running and can be used. In the integration cloud, you will see now that the agent is running.
Unfortunately (currently), it is not possible to define a virtual cloud network for the OIC to create a site-to-site connection between the Oracle Cloud and the customers network, so that the OIC connectivity agent is a good way to access on-premise systems that are not exposed to the internet. It is important to know the limitations of the adapters together with the agent. We suggest an installation of the connectivity agent on a Linux server.
Oracle recently released E-Business Suite Cloud Manager Release 22.2.1. There is an official blog post on the new features, but let’s have a look at my experience while migrating to this latest version for one of our customers.
Basically there are three areas of new features:
Let’s have a look at all of them:
In an environment with multiple application tiers (as it is common with e.g. external nodes or with larger, highly available environments) in a “shared filesystem” configuration previously data was residing on a block volume attached to the primary node. From there the /u01 directory was exported as an NFS share to the additional application server nodes. With the latest version of Cloud Manager it is possible to store the application server files of the E-Business Suite apps tier onto an OCI File System Storage Service (FSS). Oracle Cloud Infrastructure File Storage service provides a durable, scalable, secure, enterprise-grade network file system.
This has multiple advantages:
However unfortunately there are also some drawbacks of this update:
I’ll look into the exact procedure in a later blog post for sure.
Basically there are two ways to create Production2Test copies using Cloud Manager:
Pro’s:
Con's:
Pro’s:
Con's:
With the latest version of Cloud Manager one of the main drawbacks that forced me to use the “Backup and Restore” approach with many of my customers has gone away. Reason enough to give this a try.
While trying out the new clone feature my first attempts were failing with some strange error of “Failed at Update Database context file vars”. With the help of Oracle Product Management we were able to find the reason for this by changing the ruby-cookbook (~/apps-unlimited-ebs/resources/chef-repo/cookbooks/oracle-ebs/recipes/computeclone-db-rename-config.rb) to not hide sensitve stuff (comment out the line):
Then upload the modified cookbook(s) using:
knife cookbook upload -a -V -o ~/apps-unlimited-ebs/resources/chef-repo/cookbooks
This gave us an error-message (still cryptic) and finally pointed us into the right direction. While creating the instance we had to choose the new sys/system password:
There is a note to have 2 special characters which must be _ # or -. However it better be not #. That is interpreted as a comment in the bash command and then leads to a fail. With a different password it is working.
I’ve created a bug with Oracle Development and I’m sure they’ll fix this, until then better go for _ or – as the special character and not put it at the beginning or the end of the password just to be sure.
Unfortunately that brought us right away to a second issue: Since we were already on AD/TXK Release 13 with the preparations for a system-less DBA we ran into a second bug: The tooling does change the system password of the target clone to the value specified in above screenshot; but it does NOT update the EBS_SYSTEM password. However, various autoconfig-related operations then try to connect to EBS_SYSTEM with the wrong password which obviously fails. The workaround in the end for now is to NOT change the SYS/SYSTEM/EBS_SYSTEM password as part of the cloning and eventually do this as a post-cloning operation.
E-Business Suite Cloud manager in the latest version is an even better way to manage your E-Business Suite environment in Oracle Cloud Infrastructure (OCI). With the latest cloning improvements it is finally possible to create a Production-To-Test Copy usable e.g. for error diagnostics within less than an hour (even with just a single OCPU as the target) and just minutes of effort:
