4 Replies Latest reply on Oct 17, 2019 7:02 AM by Mark Francome

    DR question

    Jeffrey Jordan
      Share This:

      I was thinking about our DR plan if we lost one of our controlm servers & had to rebuild it. As far as the controlm tables/folders are they part of the database or is there a file that contains all that info similar to the accounts.xml file that has all the AFT connections info. Running 9.0.19. I know I can export all the tables from my instance & back that file up somewhere but don't want to go to that length if I don't have to. We do run filesystem as well as SQL server database backups of the server through Commvault so those would be the basis of rebuilding the server.

        • 1. Re: DR question
          Mark Francome

          For the scheduling tables and calendars, everything is in them (as you point out though, module account info is kept locally on all the Agents). There are several "system sided" elements that you should consider, e.g. quantitative resources (I backup the details daily to a flat file) but if you are doing backups as described then you should be ok (be careful with those account details on the Agents though, that can be painful).

           

          Speaking of DR in general, BMC added a lot of new DR functionality with 9.0.18 - it might be a good idea to check the new features out (especially as you are on 9.0.19).

          1 of 1 people found this helpful
          • 2. Re: DR question
            Keyur Patel

            Hi,

             

            The way I approach DR is, having Control-M Database on separate servers then Control-M EM and Control-M Server, Setup Control-M High availability between Primary and Secondary Control-M EM and Control-MM Server

             

            if you using SQL server for Control-M database then you can setup SQL replication between two DBs between Primary and Secondary DB servers

             

            Also, I would suggest to take a backup of Control-M EM and Control-M server database every day before and after New day process, so you can restore database from the backup in the event of Rebuilding DB servers

             

            This only high level, based on your Infrastructure setup you have to decide best strategy which fits for your requirements

             

            Hope this will help

             

            Thanks

            1 of 1 people found this helpful
            • 3. Re: DR question
              Jeffrey Jordan

              We run our prod/dev instances on different servers but all aspects of either instance (EM/CM/DB) are on the server, we don't really promote things from dev to prod though we do clone tables/jobs over going one way or the other so tablewise 2 totally different things plus prod/dev tables rarely match each other for changes (single oracle/banner prod instance running on prod controlm server but multiple oracle/banner dev instances running on dev controlm server). In our old solaris days we used to run mirroring servers but did away w/ that as too expensive to maintain (decision made above my pay grade). With how easy it is to do vm's vs having physical box these days, I'm thinking of trying to promote a offline mirrored prod server at least if not both prod/dev to the decision makers. Right now our DR plan is to spit out a same specs bare bones blank slate vm then restore from backups so trying to figure out what I'd get if we actually did that so to figure out the pitfalls to fill in the gaps.

              • 4. Re: DR question
                Mark Francome

                Previously I let the platform do the work; we had HP-UX and used MC/ServiceGuard to handle failover. It was just about perfect, the servers could failover and it was only a matter of a few seconds delay in processing. The worst thing that we saw was the EM clients would take a little longer to reconnect, but it was still just a number of seconds and often users were unaware of this (apart from that pesky pop-up box that said the client connectivity was lost).

                 

                Back in the day the Control-M "built-in" mirroring actually required manual intervention, which wasn't ideal. Now BMC have added in some very nice DR features (e.g. 9.0.19) and I would consider using the features shipped in the box to handle DR (seeing as they are free to use).

                 

                And a word of caution on the DB side, we used an external Oracle DB and then went to a big virtualized environment. Before doing so we asked Oracle if this impacted our license, they said no. After the project they did an audit and decided that it did impact our license and failing across would be viewed as using a brand new server, for which we would then be charged. At this point we moved all the Oracle DBs back onto plain physical machines. Beware the Oracle!

                1 of 1 people found this helpful