Tag Archives: install

Getting Started with OEM 13cR2 (13.2.2.0.0) Enterprise Manager

oem13c_mainpage_ss
OEM13c Release 2 Enterprise Manager Summary Page

Relatively easy installation, a new learning curve to master for post-installation maintenance and patching.
The principal changes in Release 2 of 13c are the segregation of the Cloud Management Services plugins (and related functionality) to specific licensable plugin components – namely:

  • Cloud Services Management – used for provisioning, allocation, and administration (think “infrastructure”) of Oracle Public Cloud-based resources (databases, development tools, middleware, etc.)
  • Oracle Cloud Application – a portal punch-out to access Oracle Cloud-based applications such as hosted e-Business Suite, Cloud Content Management, Cloud SOA Suite, Cloud Identity Management, etc. (think “apps”.)
  • Cloud Framework – a repackaging of the EM OMS framework components, and the only portion of the “cloud” components included in the base OEM licensing.
    Installation guide: http://www.oracle.com/technetwork/oem/grid-control/documentation/oem-091904.html
Oracle Cloud Framework
Cloud Management related plugins for OEM13cR2

This proof-of-concept testbed system was based upon the latest available installer set – Oracle Enterprise Manager Cloud Control 13c Release 2 Plug-in Update 1 (13.2.0.0) – we happen to be testing backwards-compatibility with some 10g databases before committing to the upgrade.  Direct (out-of-place) upgrades are supported for version of EM 12.1.0.4 and newer.

Download page for OEM 13c R2
http://www.oracle.com/technetwork/oem/enterprise-manager/downloads

Fusion Middleware in this release is Weblogic 12.1.3
The host system is built upon Windows Server 2012 R2, single-tier, 16GB RAM, 2 cores, also hosting the OMS repository database in a 12.2.0.1.0 EE installation.  Since the DB Templates provided seemed to be specific for a 12.1.x database (and errors out in certain steps during the build), I simply created a generic database from the standard template with the pre-requisite parameter settings (e.g. _allow_insert_with_update_check=TRUE; sessions > 300; shared_pool_size (10% of SGA); etc.) , and pre-created the required tablespaces:

MGMT_AD4J_TS
MGMT_ECM_DEPOT_TS
MGMT_TABLESPACE

  During the installation, the (Oracle Universal Installer) OUI will create the following new users assigned to these tablespaces:

Open:
SYSMAN
SYSMAN_TYPES
SYSMAN_BIPLATFORM
SYSMAN_STB
SYSMAN_OPSS
SYSMAN_MDS
MGMT_VIEW

Expired & Locked:
CLOUD_ENGINE_USER
CLOUD_SWLIB_USER
EUS_ENGINE_USER
SYSMAN_RO

The standard download set includes a single executable file (.exe or .bin) and several additional ZIP archives.  These are to be staged in a single directory, and the ZIP files do NOT need to be pre-extracted.

Overall, the installation was straight-forward and as-documented (despite this being a Windows installation – which usually has its own quirks.) The portion that I feel has the steepest learning curve is post-installation maintenance, and particularly patching. The new tool in 13c, the OMSpatcher (Patchset 19999993 – think of it as a Java wrapper for opatch) was really designed for use specficially in a cloud-based environment and hasn’t been completely polished for use on-premise in smaller installations.

For example, it is designed to rely upon an active Weblogic AdminServer to instantiate it’s requests for inventory versioning information and component availability for patching. But if any of those components are actually in-use by the AdminServer itself, it will tend to fail to apply, and instead provide a lengthy step-by-step instruction set of how to accomplish all of the steps manually. Depending on the complexity of your OEM setup (e.g. the number of registered plugins and target types) this could take over an hour to get the patching utility to fail and then provide the manual steps.

The OEM-specific version of OPatch (Patchset 6880880) is quite different than its predecessors.  While the OMSPatcher uses the prior “unzip into the existing Oracle Home” installation technique, the OPatch utility itself now is a Java installer that relies on the OUI to allow updating of component versioning and inventory adjustments. Most importantly, you will need to “install” the new version of OPatch before being allowed to continue with your planned patching:

$JAVA_PATH\java -jar $PATCH_STAGE\6880880\opatch_generic.jar -silent oracle_home=$ORACLE_HOME (being updated for OPatch)

This OPatch update needs to be applied (before patching) to every Oracle Home using 13c technology (the OMS home, the Agent Home, other FMW homes at version 12.1.3 or newer, etc.)

See MOS EM 13c: How to Apply a Patch to the Enterprise Manager 13c Cloud Control OMS Oracle Home (Doc ID 2091619.1) for details about making an OMSPatcher property file, and why you want to create one.

Final steps executed:
Build general purpose 12.2.0.1.0 database with new tablespaces.
Install Oracle Enterprise Manager Cloud Control 13c Release 2 Plug-in Update 1 (13.2.0.0)
Patch 19999993: EM OMSPatcher latest version 13.8.0.0.2 (unzip to ORACLE_HOME)
Patch 6880880: EM OPatch latest version 13.9.0.0.0 (both OMS and Agent)

cd $PATCH_STAGE\<patch#>
$FMW_HOME\OMSPatcher\omspatcher apply -analyze -property_file 
$FMW_HOME\OMSPatcher\omspatcher apply -property_file  [OMS_DISABLE_HOST_CHECK=true  -- a useful option added to deal with virtual host names]

Patch 25163555: Tracking bug for Back-porting 24588124 oms side fix
Patch 25604219: MERGE REQUEST ON TOP OF 13.2.0.0.0 FOR BUGS 25497622 25497731 25506784
Patch 25387277: APR-2017 PSU OMS 13.2.0.0.170418
Patch 25162444: EM-BEACON Bundle Patch 13.2.0.0.161231 (Agent)
Patch 25580746: EM-AGENT Bundle Patch 13.2.0.0.170331 (Agent)

Other useful references:

Enterprise Manager 13.2 Master Bundle Patch List (Doc ID 2219797.1)
13.2.0.0.170418 Enterprise Manager Base Platform Patch Set Update (PSU) Readme for Oracle Management Server (OMS) (Doc ID 2246778.1)

Advertisements

Using Demo Data for Oracle Data Mining Tools

With the forthcoming (but already available) SQLDeveloper 4.1 edition, an improved version of the Oracle Data Miner tools is incorporated into the SQLDeveloper console.  However, I found that there were a number of steps needed to actually use this new data modeling product other than just responding ‘Yes’ to the “Do you wish to enable the Data Miner Repository on this database?” prompt.

Here’s what I ended up doing to get things up and running (so that I could play with data modeling and visualization using Excel and the new SQLDeveloper DM extensions.)

#In this case, I’m adding back the demonstration data (i.e. EMP, DEPTNO type tables; the SH, OE, HR, et.al. schemas) into an existing R12 e-Business Suite (12.1.3) instance.

# Installing the Oracle Demo data in an R12 instance.

# Use the runInstaller from the R12 $ORACLE_HOME
cd $ORACLE_HOME/oui/bin
export DISPLAY=<workstation IP>:0.0
./runInstaller

# Choose the source products.xml from the staging area – Download and stage the DB Examples CD from OTN
/mnt/nfs/database/11203/examples/stage/products.xml

# Complete the OUI installation through [Finish]

cd $ORACLE_HOME/demo/schema
mkdir -p $ORACLE_HOME/demo/schema/log
echo $ORACLE_HOME/demo/schema/log/     ## used to respond to the Log Directory prompt during mksample.sql

sqlplus “/ as sysdba”
— will need passwords for: SYS/SYSTEM and APPS (used for all of the demo schemas, some of which pre-exist such as, HR, OE (PM, IX, SH and BI were okay for 12.1.3).

— ## Be sure to comment out any DROP USER <HR, OE, etc.) commands in this script (or you will be restoring your EBS instance from a backup because it just dropped your Module schema tables…) ##
— They look like this:
/*
mksample.sql:– DROP USER hr CASCADE;
mksample.sql:– DROP USER oe CASCADE;
mksample.sql:DROP USER pm CASCADE;
mksample.sql:DROP USER ix CASCADE;
mksample.sql:DROP USER sh CASCADE;
mksample.sql:DROP USER bi CASCADE;
*/

–Similarly – if/when you decide you no longer need the data – do NOT just use the $ORACLE_HOME/demo/schema/drop_sch.sql script
–or you just dropped your HR/OE/BI EBS schemas; don’t do that.
/*
drop_sch.sql:PROMPT Dropping Sample Schemas
drop_sch.sql:– DROP USER hr CASCADE;
drop_sch.sql:– DROP USER oe CASCADE;
drop_sch.sql:DROP USER pm CASCADE;
drop_sch.sql:DROP USER ix CASCADE;
drop_sch.sql:DROP USER sh CASCADE;
drop_sch.sql:DROP USER bi CASCADE;

order_entry/oe_main.sql:– Dropping the user with all its objects
order_entry/oe_main.sql:– DROP USER oe CASCADE;
order_entry/oe_main.sql:– ALTER USER oe DEFAULT TABLESPACE &tbs QUOTA UNLIMITED ON &tbs;

*/
— in this instance the $APPS_PW is synchronized to all application module schemas (i.e. AR, HR, GL, etc.)
— log directory would be the actual path from echo $ORACLE_HOME/demo/schema/log/ (including the trailing slash)
SQL> @mksample.sql

# to additionally create the Data Mining user (DM in this case)

create user &&dmuser identified by &&dmuserpwd
default tablespace &&usertblspc
temporary tablespace &&temptblspc
quota unlimited on &&usertblspc;

GRANT CREATE JOB TO &&dmuser;
GRANT CREATE MINING MODEL TO &&dmuser;       — required for creating models
GRANT CREATE PROCEDURE TO &&dmuser;
GRANT CREATE SEQUENCE TO &&dmuser;
GRANT CREATE SESSION TO &&dmuser;
GRANT CREATE SYNONYM TO &&dmuser;
GRANT CREATE TABLE TO &&dmuser;
GRANT CREATE TYPE TO &&dmuser;
GRANT CREATE VIEW TO &&dmuser;
GRANT EXECUTE ON ctxsys.ctx_ddl TO &&dmuser;
GRANT CREATE ANY DIRECTORY TO &&dmuser;
— Grant the SH Demo table and package objects to the DM user
@?/rdbms/demo/dmshgrants.sql &&dmuser

connect &&dmuser/&&dmuserpwd
— Create the Data Mining Views against the  SH Demo table and package objects
@?/rdbms/demo/dmsh.sql

@?/rdbms/demo/dmabdemo.sql — Builds the Adaptive Baynes Model demo
@?/rdbms/demo/dmaidemo.sql — Builds the Attribute Importance demo
@?/rdbms/demo/dmardemo.sql — Builds the Association Rules demo
@?/rdbms/demo/dmdtdemo.sql — Builds the Decision Tree demo
@?/rdbms/demo/dmdtxvlddemo.sql — Builds the Cross Validation demo
@?/rdbms/demo/dmglcdem.sql — Builds the Generalized Linear model demo
@?/rdbms/demo/dmglrdem.sql — Builds the General Linear Regression model demo
@?/rdbms/demo/dmhpdemo.sql — not a Data Mining program – Hierarchical Profiler
@?/rdbms/demo/dmkmdemo.sql — Builds the K-Means Clustering model demo
@?/rdbms/demo/dmnbdemo.sql — Builds the Naive Baynes Model data
@?/rdbms/demo/dmnmdemo.sql — Builds the Non-negative Matrix Factorization model
@?/rdbms/demo/dmocdemo.sql — Builds the O-Cluster model Demo
@?/rdbms/demo/dmsvcdem.sql — Builds the Support Vector Machine model demo
@?/rdbms/demo/dmsvodem.sql — Builds the One-Class Support Vector Machine model demo
@?/rdbms/demo/dmsvrdem.sql — Builds the Support Vector Regression model demo
@?/rdbms/demo/dmtxtfe.sql — Builds the Oracle Text Term Feature Extractor demo
@?/rdbms/demo/dmtxtnmf.sql — Builds the Text Mining Non-Negative Matrix Factorization model demo
@?/rdbms/demo/dmtxtsvm.sql — Builds the Text Mining Support Vector Machine model demo

## End of Data Mining Demo user (DM) setup and configuration for use of Oracle Demo Data

How to migrate EM12c R3 OMS and repository to a new host

Very useful, even if you’re moving parts of OEM 12c from one host to another, or modifying the filesystem mounts.

Pardy DBA

(EDIT 20130917: If you simply need to change the IP address of your OEM server, please review MOS note 1562029.1.  The procedure in that note may allow you to change your OEM server’s IP address without following the lengthy process I describe below.)

In order to save power in our data center, I need to migrate my EM12c R3 environment from the host where it currently runs to a new host.  I have a simple configuration, with a single OMS, no load balancer, and the repository database runs on the same host as EM12c R3 itself.  I also have BI Publisher installed and integrated with EM12c, and a few third party plugins as I’ve detailed elsewhere on this blog.  If you use an OS other than Linux x86-64 I suggest you research thoroughly as this procedure may or may not apply to your environment.  Further, if you have a multi-OMS…

View original post 3,770 more words

Changing ORACLE_BASE in Oracle 11gR2 (11.2.0.3.4) Database

During a routine installation, I managed to mis-type the filesystem path for the ORACLE_BASE during execution of the Oracle Universal Installer (./runInstaller).

[nb: ORACLE_BASE by normal definition should be a standard starting filesystem mountpoint under which all other instance-specific data/files for this ORACLE_HOME would be found. Typically contains your diagnostic logs, Fast Recovery Area files, all configuration tool logs, and any Flashback checkpoint saves.

$ORACLE_BASE/diag/.

./asm
./clients
./crs
./diagtool
./lsnrctl
./netcman
./ofm
./rdbms
./tnslsnr

$ORACLE_BASE/admin/<SID>/.

./adump          # Audit data./dpdump       # Data Pump related
./pfile               # init.ora files generated by config tools

–end of note]

Since the service account (oracle) had filesystem permissions to do so, it went ahead and put all of the flashback recovery, diagnostics, and audit trail files (and my backup controlfiles) into the incorrectly named directory.  The database itself was running, but how to correct this without starting all over again?

Start with backing up your init.ora configuration from spfile to pfile format:

SQL> create pfile from spfile;

File created.

(this defaults to $ORACLE_HOME/dbs/init.ora)

Then shutdown services:

SQL> shutdown immediate;

Database closed.

SQL> exit;

#oracle> lsnrctl stop

Now go find all the places you’ll need to modify/correct the ORACLE_BASE filepath declarations:

$ORACLE_HOME/dbs/init<SID>.ora

$ORACLE_HOME/network/admin (or $TNS_ADMIN)

  • listener.ora: ADR_BASE_LISTENER
  • sqlnet.ora: ADR_BASE
    (Note: these two are auto-generated using netca – the Oracle Net Configuration Assistant, if you execute that after making the database and environment-side changes.

$ORACLE_HOME/inventory/ContentsXML/oraclehomeproperties.xml : ORACLE_BASE

All finished editing? Start everything back up:

#oracle> lsnrctl start

SQL> startup pfile=?/dbs/init<SID>.ora  — to use revised settings

SQL> create spfile from pfile;       — copy new settings to spfile

SQL> shutdown immediate;

SQL> startup;                                — uses spfile instead

Now future installs using dbca, dbua and netca should recognize your defined/modified/corrected ORACLE_BASE.