Share:|

Need for High Availability? Interested in reducing downtime? Want to ensure application continuity?

 

On Wednesday, March 30th, Martin deCastongrene demonstrated, step-by-step, how you can successfully implement high availability for your Control-M/Enterprise Manager 9. During this Connect with Control-M webinar, Martin presented and demonstrated the following:

 

Provide an overview of the new Control-M/Enterprise Manager 9 High Availability
Give step-by-step installation guidance for the Control-M/Enterprise Manager 9 High Availability failover node
Show how to configure Control-M/Enterprise Manager 9 for High Availability

 

Here is the Q&A for this webinar (Connect With Control-M: High Availability with Control-M/Enterprise Manager 9)

 

________________________________________________________________

 

Q: When will this capability be available for PostgreSQL?

A: The integration of PostgreSQL for Control-M/Enterprise Manager is being looked at and will be available in a future Control-M/Enterprise Manager Version or FixPack.

________________________________________________________________

 

Q: How are extra items such as Forecast, BIM, Self Service etc. handled?  Do they need to be installed on the HA server also?

A: The additional installations such as BIM, Forecast and Self Service are activated when installed on the Primary Control-M EM  platform.  As the Secondary installation points to the same database as on the Primary, no additional action is needed on the Secondary.

________________________________________________________________

 

Q: We would like to have the failover instance in a remote datacenter. Are there any latency limitations? What if the database did not failover?

A: DNS and network latency will need to be responsive.  Testing should show if it is successful.

   The Secondary Node will failover to the same Database the Primary Node originally uses.

   If the failover does not complete it will automatically fall back to the Primary node.

________________________________________________________________

 

Q: Can a previous standalone Control-M/EM installation be converted to a secondary HA failover node?

A: For a standalone to be converted, it will need to be uninstalled and reinstalled as a High Availability Secondary Node.

________________________________________________________________

 

Q: Are the Failover Dialog items logged in a file somewhere that could be viewed at another time?

A: The EM CHA (Coordination for High Availability) process can be set to a higher debug level (MiniumDbgLvl 5 5) in order to log failover events to the cha*.log file located under the Control-M EM installation "log" directory

  ________________________________________________________________

 

Q: How long should the failover process normally take?

A: Anywhere in the 3-5 minute range would be considered normal.

________________________________________________________________

 

Q: Why isn't PostgreSQL database supported for HA in EM?

A: Development is continuing work on integrating a PostgreSQL solution for HA.

________________________________________________________________

 

Q: We are planning to set HA on our DR datacenter, can we install database on both sites (primary and DR)?

A: As there is only one Database, it is recommended to be on a third machine, which can be located in either the network where the Primary is or in DR.  The Database just needs to have good network response time to and from the Primary Node and the Secondary Node.

________________________________________________________________

 

Q: The switch over looks great out in demo, what are the common issues which we may come across or based on your experience any comments on the same...

A: DNS and network latency should both be performing well for the Failover and Fallback to complete.

________________________________________________________________

 

Q: Can you point the HA instance to an alternate Oracle DB that is kept in sync with the primary via Oracle Golden gate?

A: The Secondary HA instance is designed to point to the Primary database. We would ask to open a support ticket to discuss specific requirements.

________________________________________________________________

 

Q: Notice webinar did not actually go through HA secondary install or show HA status of primary/secondary after install. I found that the Secondary EM and CTM config agent is not auto started either after install or secondary reboot.

A: The Control-M Install Guide has a section on what to include in startup scripts for Enterprise Manager on the Primary to start.

   Only the start_config_agent script section will be needed for the Secondary startup.

________________________________________________________________

 

Q: How can I leverage Control-M failover to have 24x7 availability when applying maintenance?

A: You can Failover from the Primary to the Secondary and run on the Secondary while applying OS level maintenance or reboot.

   After Fallback to the Primary the same can be done on the Secondary.

________________________________________________________________

 

Q: In what order do I have to apply fixpacks (primary first, then secondary)?

A: EM Fixpacks are to be applied first to the Primary and then to the Secondary HA platform.

________________________________________________________________

 

Q: When the failover occurs, will the root_menu on the primary be able to manage the secondary also?

A: Once the Failover occurs to the Secondary platform, the root_menu on the Secondary must be used to manage itself.

________________________________________________________________

 

Q: You have to manually start the EM and CTM config agent and put startup scripts in place to start after reboot of secondary. This needs to be documented in Admin and/or Install guides.

A: The Control-M Install Guide has a section on what to include in startup scripts for Enterprise Manager on the Primary to start.

   Only the start_config_agent script section will be needed for the Secondary startup.

________________________________________________________________

 

Q: Can you use the command line to initiate a failover, rather than using the ccm interface?

A: You can use the stop_config_agent script to shutdown the Configuration Agent and the CHA and that will trigger a Failover.

________________________________________________________________

 

Q: Where does Oracle db reside on primary or secondary ?

A: Our recommendation is that a third machine be used for the Oracle Database.

________________________________________________________________

 

Q: Any plans to create a similar interface in the CCM for failing over Control-M Server?

A: The Control-M Server HA is also supported with the High Availability Tab in the Control-M Configuration Manager.

________________________________________________________________

 

Q: When you say that Fix Packs need to be applied to both nodes and then restarted before Failover and Failback can be accomplished, what (exactly) are the steps for applying a fix pack to both nodes?

A: You will want to bring both Primary and Secondary nodes down and EM Fixpacks are to be applied first to the Primary and then to the Secondary HA platform.

________________________________________________________________

 

Q: Will there be Failover/High Availability for the database as in version 7.0

A: The Control-M Server still can enable Mirroring in version 9.

_______________________________________________________________

 

Q: I have noticed in version 8 if I get disconnected I have to logout and back on to get a solid connection. Will the failover process have the same issue?

A: After Failover or Fallback you may need to press the retry button on the connection Dialog.

________________________________________________________________

 

Q: How long does it take automatic failover to happen?

A: Anywhere in the 4-6 minute range would be considered normal.

________________________________________________________________

 

Q: The presentation says: Fixpacks will need to be applied on both nodes... Can we failover to the secondary, apply FP2, and then fail back to primary, and apply FP2 on the secondary?

A: You will want to bring both Primary and Secondary nodes down and EM Fixpacks are to be applied first to the Primary and then to the Secondary HA platform.

________________________________________________________________

 

Q: Is there an additional license charge for HA?

A: You will want to discuss license issues with your Account Manager.

________________________________________________________________

 

Q: In this scenario, you assume 2 em on 2 different hostnames and 1 database in 1 some different hostname. What if the database is down or fails? is there some scenario of ha that includes the database?

A: Please open a support ticket and we can review the various options available with high Availability and disaster recovery as it relates to the database.

________________________________________________________________

 

Q: We have set HA with Postgres, should we enable archive mode for the PostgreSQL database?

A: You are welcome to enable the archive mode for the PostgreSQL.

________________________________________________________________

 

Q: If the alternate primary option is activated for a secondary node, is it possible to rollback the primary for the primary node without reinstall the first node?

A: Yes, you can just select the Fallback to Primary option and it will return to the original configuration; running on the Primary.

________________________________________________________________

 

Q: I found that the secondary EM and CTM config agent is not auto started either after install or secondary reboot. You must manually restart and create scripts for reboot.

A: The Control-M Install Guide has a section on what to include in startup scripts for Enterprise Manager on the Primary to start.

   Only the start_config_agent script section will be needed for the Secondary startup.

________________________________________________________________

 

Q: Do we need a shared disk in cluster for the installation?

A: The Enterprise Manager Database and Primary Enterprise Manager and Secondary Enterprise Manager can be on a standard disk with no cluster needed.

________________________________________________________________

 

Q: When it's necessary to debug this feature, should we enable debug on both sides, primary and secondary?

A: You will want to enable debug on both sides, Primary and Secondary.

________________________________________________________________

 

Q: Is there a way to configure the EM API web service to HA?

A: Currently High Availability is being supported for the Control-M/Enterprise Manager and Control-M/Server only.

________________________________________________________________

 

Q: Can you define "same database"? SQL Listener is a single point of contact for Always On multi node setup. Will EM work point to the non-changing listener?

A: Please open a support ticket and we will gladly review the Oracle scenario in your environment.

________________________________________________________________

 

Q: If you weren't logged in the CCM, and it failed over, will the banner be displayed once I log in?

A: The Failover banner will be displayed the next time someone does login to the Control-M Configuration Manager

________________________________________________________________

 

Q: When the HOST from primary is DOWN, the CLIENT connection remains asking for the primary node name for connections on the Secondary with the HA Secondary ACTIVATED?

A: This question has been answered verbally.

________________________________________________________________

 

Q: If I have 2 environments setup with HA and then mirror the DB to the 2nd environment  if I shut down the 1st environment and bring up the 2nd can I expect that to work?

A: Please open a support ticket and we will gladly review your environment.

________________________________________________________________

 

Q: What will be the status of the executing jobs in the monitor? (during the failover process)

A: During the failover the Monitoring will be suspended until the GUI Server and Gateway have started on the Secondary node and the WLA Client is reconnected.

________________________________________________________________

 

Q: Does Control_M record the reason for failover and log the same ?

A: The EM CHA (Coordinator for High Availability) process can be set to a higher debug level (MiniumDbgLvl 5 5) in order to log failover events to the cha*.log file located under the Control-M EM installation "log" directory

________________________________________________________________

 

Q: If a FP contains DB changes then you can't failover correct?  You will have to shut down both EMs?

A: Correct.  You will want to bring both Primary and Secondary nodes down and EM Fixpacks are to be applied first to the Primary and then to the Secondary HA platform.

________________________________________________________________

 

Q: When failing over could there be any issues getting Alerts or Global conditions during the fail over process?

A: During the actual Failover Alerts and Global Conditions will be suspended until Failover completes.  Then any queued up Alerts or Global conditions will be propagated.

________________________________________________________________

 

Q: How would the account management settings change for Control-M Web Services API ?

A: This question has been answered verbally.

________________________________________________________________

 

Q: For Control-M environments with Mainframe and Distributed are there any mainframe parameters that have to be configured for the IOAGATE for HA?

A: The Gateways for the Control-M datacenters, whether Mainframe or Distributed are defined and activated initially on the Primary platform and require no special handling when moved to the Secondary.

________________________________________________________________

 

Q: Can I automate, or schedule when to execute a failover from primary to secondary?

A: Using the stop_config_agent script with the password file option it could be possible to automate the failover.

________________________________________________________________

 

Q: Will HA work if the primary and secondary are at different fixpacks?

A: The Primary and Secondary Nodes need to be at the same FixPack level.

________________________________________________________________

 

Q: The install of the HA is part of the product install, I assume. Or is it a separate install?

A: The HA functionality is part of the Base install for Control-M EM 9.0.00, there isn't a separate download. We strongly recommend applying the latest fixpack to ensure you have the most recent fixes on the system.

________________________________________________________________

 

Q: Clarification for the FP install. If the FP we are installing changed the database, will the secondary (still not updated) know that it needs to operate with a possibly different schema?

A: Correct.  The FixPack install detects that it is being installed on the Secondary and installs only what is needed on the Secondary.

_______________________________________________________________

 

Q: So, in order to KNOW why it failed over DEBUG has to be turned on in order for that to be logged.  Is there not a performance impact to constantly running with DEBUG on?

A: Leaving the EM CHA process in debug may have a slight resource overhead on the system as the activity is written to the log file.

________________________________________________________________

 

Q: What processes answer for a connection from a CLIENT calling for the primary NODE when it was jumped for the secondary node? If the HOST (machine) is offline, how the connections calling for the first node (that is offline) arrives for the secondary node?

A: Please open a support ticket and we will gladly review specifics of the Failover architecture.

________________________________________________________________

 

Q: What will be the status of the executing jobs in the monitor? (during the failover process)

A: The status of executing jobs will stay in executing state until Failover is complete and the Client reconnects and the status is updated.

________________________________________________________________

 

Q: What are all the conditions need to checked before by control-m before switching to secondary EM?

A: The Primary CHA checks the Secondary CHA process in up, communicating and communicating with the Database.

________________________________________________________________

 

Q: Are the primary and secondary using the same CCM?

A: Yes.  A Control-M Configuration Manager connected to the Primary before Failover will be reconnected to the Secondary after Failover.

________________________________________________________________

 

Q: About the FPs: Then, the HA is not for applying CTM FPs. All components must be down in primary and secondary when applying FPs.

A: That is correct, when applying fixpacks, Control-M must be down.

________________________________________________________________