January 13th, 2015 by dbakevlar

Every job comes with tasks that no one likes to perform and database administration is no exception.  Patching is one of those necessary tasks that must be performed and when we are expected to do more with less everyday, the demands of patching another host, another agent, another application is often a task that no one looks forward to.  It’s not that it goes wrong, but that it’s just tedious and many DBAs know there are a lot of other tasks that could be better use of their time.   Patching is still an essential and important task that must be performed, we all know that. OPatch and other patching utilities from Oracle make patching easy, but it can still remove a lot of time from a resource’s day.

Enterprise Manager 12c’s automated patching and provisioning, using the Database Lifecycle Management Pack is gaining more appreciation from the IT community as it assists the DBA with features to search recommended patches, create patch plans, review for conflicts and allow sharing and re-use of patch plans.

Configuring a Database for Online or Offline Patching

After logging into a target database, you can click on Setup and go to the Offline Patching setup:

patching22

You can then choose to use Online patching with MOS credentials:

patching1

or use Offline Credentials and configure the patching catalog and ensure you upload all the XML’s for the catalog, which will now be stored locally to a workstation.  Once the upload is complete, run the Refresh From My Oracle Support job.

patching2

The Online configuration is recommended and works with the software library.  It’s what we’ll be talking about today.

Also ensure that you’ve set up correct privileges to perform patching. Provisioning and patching require steps to be performed that will require privileges to run root scripts, so ensure that the credentials that are used for the patching allow to sudo to root or PBrun.

Database Patch Plans

To set up a patch plan for a database, there are a number of steps, but the patch plan wizard makes this very easy to do.  For our example, we’ll choose to patch 11.2.0.4 databases to the latest recommended patches.

First, let’s do a search to find out what patches we’ll need to apply to our 11.2.0.4 databases in our EM environment.

patching3

Our Enterprise menu takes us to the Provisioning and Patching, Patches and Updates.

From this console page, we can view what patch plans are already created in case we can reuse one:

patching4

As there isn’t an existing plan that fits what we need to do, we are going to first search for what patches are recommended with the Recommended Patch Advisor:

patching10

We’ve chosen to perform a search for recommended patches for 11.2.0.4.0 databases on Linux x86-64.  This will return the following four patches:

patching11

We can click on the first Patch Name, which will take us to the patch information, including what bugs are addressed in this patch, along with the option to download or create a patch plan.  For the purpose of this post, we’ll choose to create a patch plan:

patching12

We’ll create a new patch plan for this, as our existing ones currently do not include an 11g database patch plan that would be feasible to add to.  We can see our list of patches on the left, too, so this helps as we proceed to build onto our patch plans.

After clicking on the Add to New, we come to the following:

patching13

Name your patch plan something meaningful, (I choose to name the patch for a single instance, “SI”, the patch number and that it’s for 11.2.0.4) and then choose the database from the list you wish to apply the patch to.  You can hold down the CTRL key and choose more than one database and when finished, click on Create Plan.

The patch plan wizard will then check to see if any other targets monitored by Cloud Control will be impacted and asks you to either add them to the patch plan or to cancel the patch plan for further investigation:

patching14

If you are satisfied to with the additions, you can click on Add All to Plan to proceed.  The wizard then checks for any conflicts by the additions and will report them:

patching15

In our example above, I’ve added an 11.2.0.3 instance home to show that the wizard notes it and offers to either ignore the warnings and add it or (more appropriately) cancel the patch plan and correct the mistake.

Adding to Patch Plans

In our recommended patch list, we had four recommended patches.  Once we’ve created our first patch plan, we can now choose to add to it with the subsequent patches from the list:

patching16

This allows us to create one patch plan for all four patches and EM will apply them in the proper order as part of the patch deployment process.

Patch Plan Review and Deploy

One a patch plan is created, the next step is to review and deploy it.  Choose the patch plan from the list that we created earlier:

patching18

Double clicking on it will bring up the validation warning if any exist:

patching17

We can then analyze the validations required and correct any open issues as we review the patch plan and correct them before deploying:

patching29

We can see in the above checks, that we are missing credentials required for our patches to be successful.  These can now be set by clicking to the right of the “Not Set” and proceed with the review of our patch plan.

patching20

Next we add any special scripts that are required, (none here…) any notification on the patching process so we aren’t in the dark while the patch is being applied, rollback options and conflicts checks.

These steps give the database administrator a true sense of comfort that allows them to automate, yet have notifications and options that they would choose if they were running the patch interactively.

Once satisfied with the plan, choose the Deploy button and your patch is now ready to scheduled.

patching21

Once the patching job completes or if it experiences an issue and results in executing the logic placed in the above conflict/rollback steps, the DBA can view the output log to see what issues have occurred before correcting and rescheduling.

Output Log 
Step is being run by operating system user : 'ptch_em_user' 
 
Run privilege of the step is : Normal  

This is Provisioning Executor Script
…
Directive Type is SUB_Perl
…
The output of the directive is:
…
Tue Jan 6 00:15:40 2015 - Found the metadata files; '19121551' is an patch
…
Tue Jan 6 00:15:40 2015 - OPatch from '/u01/app/oracle/product/11.2.0/dbhome_1/OPatch/opatch.pl' 
  will be used to apply the Interim Patch.
…
Tue Jan 6 00:15:52 2015 - Invoking OPatch 11.2.0.4.7
…
Following patches will be rolled back from Oracle Home on application of the patches in the given list :
   4612895
…
Do you want to proceed? [y|n]
Y (auto-answered by -silent)
User Responded with: Y
OPatch continues with these patches:  6458921  

Do you want to proceed? [y|n]
Y (auto-answered by -silent)
User Responded with: Y

Running prerequisite checks...

This is high level, but really, it’s quite easy and the more you automate provisioning and patching, the easier it’ll get and you’ll wonder why you waited so long!

 

 

Posted in Cloud, Enterprise Manager Tagged with: , ,

September 4th, 2013 by Kyle Hailey



Oracle OEM 12c introduces a new feature that enables the creation of Oracle database thin clones by leveraging file system snapshot technologies from either  ZFS  or Netapp.  The OEM adds a graphic interface to the process of making database thin clones. The feature that enables database thin cloning in OEM is called Snap Clone and is part of OEM’s Cloud Control Self Service for data cloning. Snap Clone is available via the feature Database as a Service (DBaaS). Snap clone leverages the copy on write technologies available in some storage systems for database cloning.  Support is initially available for NAS storage and specifically on  ZFS Storage  and NetApp Storage.

In order to use Snap Clone, one has to install the source database such that the source database data files are on a ZFS storage  or Netapp array and have the storage managed by agents on a LINUX machine and then one can thin clone data files on that same storage array.

Snap Clone offers role based access, so storage admin can login in and only have access to areas they are responsible for as well as limiting access to source databases, clones and resource by end users.

Setting Snap Clone

 

The prerequisites for getting start with Snap Clone are having available storage on ZFS Storage Appliance or Netapp storage array as well as having access to a master test database. A master test databse is a database that has a sanatized version of a production database such that it is either a subset and or  masked. The test master database has to be registered with OEM.  After the test master is registered with OEM, Snap Clone can be setup. To set up snap clone, come into Oracle Cloud Control 12c as “cloud administrator” role with “storage adminstator” priviledge or super administrator and register the storage. To register the storage navigate to “ setup -> provisining patching -> storage registration”.

  • Navigate to “ setup -> provisining patching -> storage registration”
  • Click “Register” tab, and choose storage, either Netapp or ZFS,
    • Supply storage information
      • Name: Storage array name registered in DNS
      • Vendor
      • Protocol: http or https
      • Storage Credentials: credentials for interacting with storage
  • Install agents on a separate LINUX machine to manage the Netapp or ZFS storage. An agent has to run on Linux host to manage the  storage. Supply the
    • Agent host
    • Host credentials
  • Pick a database to make the test master
    • Put the test master on ZFS storage or Netapp storage
    • Register the ZFS storage or Netapp storage with OEM
    • Enable Snap Clone for the  test master database
  • Set up a zone – set max CPU and Memory for a set of hosts and the roles that can see these zones
  • Set up a pool – a pool is a set of machines where databases can be provisioned
  • Set up a profile – a source database that can be used for thin cloning
  • Set up a service template – reference values such as a init.ora for database to be created

Screen Shot 2013-06-09 at 10.08.47 PM

Figure 1. Shows the entry page in OEM 12c Cloud Control. From here go to the top right and choose setup, then provisining patching then storage registration as shown above.

 Navigate to storage registration

To setup Snap Clone navigate to storage registyratiom choose the menus “setup -> provisining patching -> storage registration”.

Screen Shot 2013-06-09 at 10.10.44 PM

Figure 2 Shows a zoom into the menus to choose

Screen Shot 2013-05-31 at 10.09.02 PM

Figure 3. Storage Registration Page

Screen Shot 2013-06-09 at 10.15.04 PM

Figure 4. Choose the type of storage array

Once on the Storage Registration page, choose “Register” and then choose the storage, either Netapp or Sun ZFS.

Register the Storage

Screen Shot 2013-06-09 at 10.17.02 PM

Figure 5. Storage Registration Page

To register the storage supply the following information

  • Name: Storage array name registered in DNS
  • Vendor
  • Protocol: http or https
  • Storage Credentials: credentials for interacting with storage

All of which is documented in cloud administration guide.

Define agents used to manage storage

Then define agents used to manage storage. Agents have are required to run on a LINUX host.  More than one agen can define to provide redundancy.  The agents will be the path by which OEM communicates with the storage. For each agent, supply the following information

  • Host name
  • Credential type
  • Credentials

And finally define the frequency with which the agent synchronizes with the storage to gather the sorage hardware details such as  information on aggregates shares volumes.

After the storage information, agent information and agent synchronization information has been filled out, then hit “submit” button in the top right. Hitting the submit button will return the UI back to the “Storage Registration”. On the “Storage Registration”, click on the storage appliance listed in the top, then click on the contents tab on the bottom half of the page. This will list all the  volumns and aggregates in the storage appliance.

Looking at volumns on Storage Array

Screen Shot 2013-06-09 at 10.18.48 PM

Figure 6. Editing storage ceiling by clicking on a aggregate and hitting the “Edit Storage Config” tab.

For each aggregate one can set storage ceilings. Click on the aggregate or FlexVol and the click “Edit Storage Ceilings” tab.

Choosing a Database Test Master

On the database tab is a list of databases that can be used for cloning. OEM detects the database automatically on the hosts it is managing. OEM will also automatically correlate databases that have storage on the storage array added storage registration.  OEM looks for all databases that have files on the registered storage.  Click on database, then the show files tabs which will show the files and volumes for this database.

Screen Shot 2013-06-09 at 10.20.00 PM

Figure 7. List of files by volumn for database

Screen Shot 2013-06-09 at 10.20.07 PM

Figure 8. Enable Snap Clone for databases that will be used as test masters.

Nominating a database as test master requires enabling snap clone. To enable snap clone for a database, click on the chosen database, then click “Enable Snap Clone” tab just above the list of databases. This will automatically validate that all the volumes are flex clone enabled (in the case of Netapp).

 

Setting up Zones

The next step is to configure zone which can be used to organize cloud resources

Choose the menu option “Enterprise -> Cloud -> Midelware and Database Home”

 

Screen Shot 2013-06-09 at 10.23.11 PM

Figure 9. Navigate first to “Cloud -> Middleware and Database Home”

Middleware and Database Cloud page

Screen Shot 2013-06-09 at 10.24.23 PM

Figure 10.  Middleware and Databawe Cloud page

Setting up a Zone

In order to see the zones defined, click on the number next to the title “Paas Infrastructure Zones”  in the top left under General Information.

Screen Shot 2013-06-09 at 10.26.24 PM

Figure 11. PaaS Infrastructure Zones

To create a zone, click the tab “Create”.

Screen Shot 2013-06-09 at 10.27.33 PM

Figure 12. first page of wizard to create a PaaS Infrastructure Zone, give a meaningful name and description of the zone and define maximum CPU utilizaiton and memory allocation.

In the first page of the “PaaS Infrastructure Zone”, give zones a meaningful name and description. Define constraints such as maximum host CPU and memory allocations.

Screen Shot 2013-06-09 at 10.28.56 PM

Figure 13. Second page of the “PaaS Infrastructure Zone” wizard, add hosts that are available to the zone.

Next define hosts that are members of the zone and provide credentials that operate across all members of this zone

Screen Shot 2013-06-09 at 10.30.11 PM

Figure 14. Third page of the “PaaS Infrastructure Zone” wizard, limit which roles can see the zone.

Next define what roles can see and access this szone. The visibiliy of the zone can be limited to a certain class of users via roles like Dev, QA etc

Screen Shot 2013-06-09 at 10.31.18 PM

Figure 15. Final review page for “PaaS Infrastructure Zone” wizard

Finally review settings and click submit

Screen Shot 2013-06-09 at 10.32.29 PM

Figure 16. Showing the Confirmation that the PaaS Infranstructure Zone has been successfully created.

 

Creating Database Pool and Profiles

The remaining steps required to enable snap clone is to create a database pools which is a collection of servers or nodes that have database software installed.  The remaining part of the setup is done by a differnet user who is the administrator for database as a service.

Log in as  DBAAS_ADMIN.

For the next part navigate to the menu “Setup -> Cloud -> Database”.

Screen Shot 2013-06-09 at 10.33.24 PM

Figure 17. Middleware and Database Cloud page

Screen Shot 2013-06-09 at 10.35.59 PM

Figure 18. Navigate to “Setup -> Cloud -> Database”.

Screen Shot 2013-06-09 at 10.36.56 PM

Figure 19. Database Cloud Self Service Portal Setup. To create a database pool choose the “Create” button in the center of the page and from the pull down, choose “For Database”.

To create a new pool click on the “Create” button in the center of the page, and chose “For Database” from the pull down menu that appears.

Screen Shot 2013-06-09 at 10.38.16 PM

Figure 20. Choose “Create -> For Database”

Screen Shot 2013-06-09 at 10.39.10 PM

Figure 21. Edit pool page. Provide a meaningful name a descrpition of the pool. Add Oracle home directories in the bottom of the page. At the very bottom of the page set a constraint on the number of databases instances that can be created in the pool. On the top right, set the host credentials.

Set

  • Name and description
  • Oracle Home
  • Maximum number of databases per host
  • Credentials

In the “Edit Pool” page, at the top left of the screen, provide a meaningful name and description for the pool. In the middle of the screen add Oracle homes that will be used for databse provisioning.  Every member of a database pool is required to be homogeneous. Homogenous requires that the platform and Oracle version is the same across all the hosts  and Oracle homes in the pool. All the Oracle installations also have to be of the same type either single instance or RAC. In the top right  add the  Oracle home provide oracle credentials and root credentials. Finally at the bottom of the page a constraint can be set on the number of database instances that can be started in this pool.

Screen Shot 2013-06-09 at 10.41.03 PM

Figure 22. Set request limits on the pool

The next page sets the request settings. The first restriction sets how far in advanced can requrest can be made. Second restricts  how long a request can be kept which is the archive retension.  After the archive retention time  the  requests will be deleted.  Finally is the request duration which is the maximum duration for which the request can be made.

Screen Shot 2013-06-09 at 10.42.12 PM

Figure 23. Set memory  and storage quotas per role for the pool. The quotas cover memory, storage, database requests and schema requests.

 

The above page  configures quotas.  Quota is allocated to each and every self service user. The quotas controls the amount fo resources users  have access to. Quotas are assigned to a role and users inherit quota values from the role. Click “Create” in the middle of the screen.

Screen Shot 2013-06-09 at 10.43.03 PM

Figure 24. Editing the quotas on a pool for a  role.

The popup dialogue has for these entries

  • Role name
  • memory GB
  • storage GB
  • number of database request

Screen Shot 2013-06-09 at 10.43.56 PM

Figure 25. Profiles and Service Templates

Profiles and service templates

.A profile is use to capture information about the source database which can then be used for provisioning.

A service template is a standardized service definition for a database configuration that is offered to the self service users. A collection of service templates forms the  service catalogue. A service template will provision databsae with or  without seed data. To capture an ideal configuration, the easist thing to do is to point at an existing database and fetch information of interest from that database.  The information from the database can be captured using a profile.

To create a profile click on the “Create” button under “Profiles”

Creating a profile

Screen Shot 2013-06-09 at 10.45.07 PM

Figure 26. Specify a reference target for a profile.

Click the magnifying glass to search for a source database.

Screen Shot 2013-06-09 at 10.46.46 PM

Figure 27. Search for a reference target database

 

Pick a refrence target by clicking on it, the click the “Select” button in bottom right.

Screen Shot 2013-06-09 at 10.47.31 PM

Figure 28. Creating a Database Provisioning Profile

To pick a database for use in thin cloning, choose the check box “Data Content” with suboption selected fro “Structured Data” with sub-option selected for “Create” with sub-option selected for “Storage Snapshot”. This  option is only enabled only when the “enable snapshot” option is enabled on the storage registration page. Disable option capture oracle home.

Provide credentials for the host machines Oracle account and for the database login.

 

The “Content Option” step is not needed for the above selections.

Screen Shot 2013-06-09 at 10.48.16 PM

Figure 29. Give the profile a meaniful name and description

 

Next provide credentials for Oracle home and Oracle databse, then provide a meaningful name for the profile  as well as a location. The profile will be userful when creating a service template.

Screen Shot 2013-06-09 at 10.49.12 PM

Figure 30. Review of create profile options.

 

Next  review the summary and click subit which will connect to storage and take snapshots of the storage


Screen Shot 2013-06-09 at 10.50.46 PM

Figure 31. Shows a zoom into the menus to choose

 

To  create a new service template choose a profile and  in this case use “thin provisioning for reference DB” profile.  Now to create a new service template click “create” and choose “for database”. Service templates are part of the service catalogue and exposed to the self service users.

Screen Shot 2013-06-09 at 10.51.42 PM

Figure 32. Provide a meaningful name and description for the service template.

 

Provide a meaningful name and description. For the rest of service template provide information about the databses that will be created from the snapshots such as providing database type, rac or single instance, for rac provide number of nodes. Provide the SID prefix to appended to the SIDs generated for the clones, provide the Domain Name and the port.

Screen Shot 2013-06-09 at 10.52.38 PM

Figure 33. Provide a storage area for writes to use.

 

The   cloning operation only creates a read only copy thus it is required to provide write space elsewhere in order to allow writing to the thin clone.

click on the edit button

click on volumne, then edit button

Screen Shot 2013-06-09 at 10.53.29 PM

Figure 34. set diretory and maximum space usage for the write location

 

Provide the mount point prefix and amount of writeable space wish to allocate

 

Users of the thin clone databses can  also be allowed to take further snapshots. These snapshots can be used as a mechinism to rollback changes, The number of thiese snapshtos can limited just below storage size section:

Screen Shot 2013-06-09 at 10.54.18 PM

Figure 35. set the number of snapshots that can be taken of a a thin clone

Screen Shot 2013-06-09 at 10.55.15 PM

Figure 36. set the initial passwords for database accounts on the thin clone.

next provide credentials for administrative accounts

  • SYS
  • SYSMAN
  • DBSMNP

for all other non-administartive accounts can choose to leave them as is or change them all to one password you can modify certain init.ora parameters for exmaple memory

Screen Shot 2013-06-09 at 10.55.59 PM

Figure 37. Modify any specific init.ora parameters for the thin clone

 

Screen Shot 2013-06-09 at 10.57.13 PM

Figure 38. Set any pre and post provision scripts to be run at the creation of a thin clone.

 

custom scripts can be provide as pre or post creation steps this can be very useful if you want to register databses with  OID or certian actions that are specific to your organization

Screen Shot 2013-06-09 at 10.58.05 PM

Figure 39. Set the zone and pool for the thin clone

Screen Shot 2013-06-09 at 10.59.07 PM

Figure 40. set the roles that can use the service template.

you associate this srvice template with a zone and a template this insures that the service template can actualy work on the  group of resources that you have identified and can limit the visibility of the service tempalte usein roles

Screen Shot 2013-06-09 at 11.00.17 PM

Figure 41. review of the service template creation requites

finally we review the summary and click submit

Creating  a Thin Clone

Screen Shot 2013-06-09 at 11.06.14 PM

Figure 42. 12c Cloud Control

Screen Shot 2013-06-09 at 11.07.06 PM

Figure 43. 12c Cloud Control Self Service Portal

Contents of the Database Cloud Self Service Potal screen

  • Left hand side
    •  Notification – any instances that are about to expire
    •   Usage
      •     databases (number provisioned out of maximum)
      •     schema services
      •     Memory
      •     Storage
  • Right side
    •   Top
      •     Databases
    • Bottom
      •     requests – requests that created the database services and the database instances

Screen Shot 2013-06-09 at 11.08.02 PM

Figure 44. To clone a database, choose the “Request” then “Database” menu.

Screen Shot 2013-06-09 at 11.08.40 PM

Figure 45. From the list choose a Self Service Template. In this case “SOEDB Service Template”

Options are

  •   RMAN backups which are full clones
  •   empty databases
  •   snap clone which are thin clones

Screen Shot 2013-06-09 at 11.09.42 PM

Figure 46. Fill out the clone Service Request

request wizard asks for

  • request name
  • select zone – collection of servers
  • select a start and end time
  • provide a user name and password, new user and password

Users do not get system access to the databases but instead get a slightly less privilege user who becomes the owner of the database

Hit Submit

Screen Shot 2013-06-09 at 11.10.25 PM

Figure 47. Shows new clone database

 

References

http://www.youtube.com/watch?v=J7fnfLS5Dxg&feature=youtu.be – setup

http://www.youtube.com/watch?v=9VK1z6nU1PU – provisioning

http://www.oracle.com/technetwork/oem/cloud-mgmt/dbaas-snap-netapp-2041775.mp4

http://www.slideshare.net/kellynpotvin/dbaas-database-as-a-service-in-a-dbas-world#!

 

Posted in cloning, em, Oracle Tagged with: , ,

  • Facebook
  • Google+
  • LinkedIn
  • Twitter