Query regarding oracle RAC configuration in Admin console and its alerting

[Saini Aditi]

Hi ,

I need to know for ORACLE RAC monitoring exactly configured in admin console.

I have deployed oracle probe on both the servers. on server 1 all the database connections have been created with respective profile.The connections are showings both the databases per connection and enabled RAC checkpoints for all.

1)Need to know same thing needs to be enabled on server 2 as well and to activate the profiles at same time.

2)Is there any way we can make a grouping/template of RAC checkpoints which we want to activate and deploy over each profile like we can do in IM.

3) alerts from this probe as its configured in admin console would be visible on alarm console or not?

 

Thanks Aditi

[Kathy Maguire]

Aditi -

 

The steps for configuring the oracle probe for RAC monitoring is as follows:

 

1.  Make sure that you have followed all of the steps specified on the oracle Preconfiguration Requirements Wiki page including setting up the tnsnames.ora file with an entry for each Oracle RAC you wish to monitor with the oracle probe.

2.  Create RAC connection profiles for each Oracle RAC that you wish to monitor making sure that you select the "Is RAC" option when creating the connection profile.  NOTE:  You do not create a profile for each node in the Oracle RAC cluster.  The probe will handle discovering these when the monitoring profile is created for each RAC connection profile you define.

3.  Enable all of the checkpoints in the RAC Common and RAC Specific templates that you wish to have the oracle probe monitor, setting up the thresholds for alarms to be generated.  Disable any checkpoints that you are not interested in monitoring.

4.  From the RAC connection profile, create a monitoring profile.  On profile creation, all the active nodes are automatically added under the profile. One of the nodes is displayed as primary. This node denotes that the data for a RAC common checkpoint is routed to this node when viewed from UMP.  Even when the node is down, the data is still routed to this node (to avoid any data loss).

Note: When you bind a profile to the RAC connection, all the RAC common and node-specific checkpoints are automatically loaded in the RAC profile. The common checkpoints are added directly in the profile and the node-specific checkpoints under the nodes.

5.  If you wish to make specific changes to any of the checkpoints applied from the templates, you can make them under the RAC profile itself.

 

Please note that it is recommended that you use the latest 5.30-T2 Oracle hot fix that should be available for download from the CA UIM Hotfix Index site in the near future.  If you cannot wait for this site to be updated with the latest 5.30-T2 oracle hot fix, then open a case to request a copy of it.

 

Hope this helps.

[Saini Aditi]

Thanks a ton Kathryn for configuration steps,its really helpful!

• In point 2 of yours, when I created the monitoring profile for a connection, both nodes got automatically discovered, but despite of activating the RAC check points node specific and RAC common, all the RAC Checkpoints were reflecting in the profile and activated one were not activated in the profile. Due to this for every profile I had to manually activate the required RAC Checkpoints.
• Also please confirm same RAC connections and monitoring profile need to be created on server 2 as well?
• How could we take backup of the required active  RAC checkpoints from one server and save it, so that we can deploy the check points on other servers and do not have to manually activate every time we get RAC monitoring , only just have to create connections and monitoring profile. Is there any method for same.
• Also we have created one RAC common checkpoint in the template, its query was working when it was created.But now after few days getting error

 

Thanks

Aditi

[Kathy Maguire]

Aditi -

 

The  5.30-T2 oracle hot fix which is currently posted to the CA UIM Hotfix Index site addresses the problem you noted where some checkpoints enabled in the RAC templates are not enabled after the templates are applied to the RAC monitoring profile.

 

If Server 2 is one of the nodes in the Oracle RAC cluster you are monitoring from Server 1, then it should appear in the RAC profile you created in the oracle probe on Server 1 as one of the RAC specific nodes

 

 

You might be able to make a backup copy of the RAC checkpoints you want to enable so that you can get these setup by default in a new oracle probe deployment as follows (this assumes you are using the 5.30-T2 oracle probe in all cases).

 

1.  From an oracle 5.30-T2 probe deployment delete all RAC profiles.

2.  Enable the checkpoints you need to monitor in the RAC Common and RAC specific templates defining thresholds that you want for the enabled checkpoints.  Disable the checkpoints you do not want to monitor.  NOTE:  Do not disable the check_dbalive checkpoint in the RAC common template.  If this checkpoint is disabled, the probe may not execute any of the other enabled checkpoints.

3.  Save the changes.

4.  Drag and drop a copy of this oracle probe from the robot to the local archive on your primary hub to create a configuration only oracle probe package - make sure your rename the package before saving it.

5.  Edit the new oracle configuration only package you created, then edit the oracle.cfx file in the package.  Remove all sections from the oracle.cfx file except the <rac_common_checkpoints> and <rac_specific_checkpoints> sections and their contents.  Save the change to the oracle.cfx file then save the modified oracle configuration only package.

6.  After deploying the 5.30-T2 probe to an new robot, deploy the oracle configuration only package then from the Admin Console create your RAC connection profile followed by the RAC profile

 

For the connection error you are now seeing from the probe, have you tried testing the connection to the configured Oracle RAC cluster?  Does the connection test succeed?  If not, you need to determine what is preventing the probe from connecting to the configured Oracle RAC cluster.  If the connection test succeeds, then you need to change the oracle probe's loglevel to 5 and logsize to 80000, apply the changes and restart the probe.  Allow the probe to run for a couple of monitoring intervals then check the logs to see if you can determine why the checkpoint is failing.  If you cannot figure this out from the logs, then open a case posting the logs and the oracle_monitor.cfg file to track down the problem.

 

Hope this helps.

[Saini Aditi]

Hi Kathryn,

Thanks for update.

Not able to downloas 5.30-T2 oracle hot fix, site cannot be reached , could you share the same here, or can i get it from some other site?

 

Not the 2 server is not one of the node we are monitoring, Its kind of pointing to the same Oracle DB nodes as server 1.(so if we want we can configure it as server 1 , in case agent goes down on server 1 , monitoring is not hampered as its active from server 2?

 

I will check and try to make Oracle rac backup in admin console once get the hotfix.

 

Also yes all the connections configured for Oracle RAC cluster are successful.But error is their for checkpoint we have created for RAC .Will check as u suggested otherwise willl check with CA support case.

 

Thanks Again for all the information.

 

Aditi

 

[sirha02]

Hi Aditi,

 

Try this link for the hot fix. The release notes can be found here

[David Michel]

If your company blocks the site or downloads from the hotfix site, open a Support case requesting a copy of the hotfix is attached to the case for you to download from there.

[Saini Aditi]

thanks for the link sirha!