6 Replies Latest reply on Feb 10, 2017 6:16 AM by phoffmann

    How do you sync device inventory across 2 core servers for Disaster Recovery?

    Apprentice

      We are currently in the process of making our DR process for LANDesk a little cleaner. We currently are using database mirroring and backups but that does not provide a clean database when swapped to the DR server (there are a lot of database artifacts that have to be cleaned up as they point to the wrong server). I have found that most everything I need to backup can be synced using Core Sync (no more database mirroring). The only piece I haven't been able to get working yet is device inventory sync. Our DR server won't do us a lot of good if it has no devices in it for a few days as devices have to send in full sync scans on a failover. I have read up on the stored scans functionality and I am thinking that may do what I need but I'm not sure - https://community.landesk.com/support/docs/DOC-27764.

       

      Our environment:

      Production - Physical core server with database on production SQL cluster

      Core server (DR) - Physical core server with database on DR SQL cluster

       

      Is there anyone that has a similar setup?

      What is the best way to keep everything in sync between the 2 servers, including device inventory?

        • 1. Re: How do you sync device inventory across 2 core servers for Disaster Recovery?
          MarXtar ITSMMVPGroup

          I understand why you are trying to do what you do, but the methods you are using really don't fully achieve what you want, nor are they supported for the purpose they are being used. By this I simply mean if you need to use the DR and it doesn't work, LD support can't support it.

           

          When asked about DR I generally ask how essential the Core and Database is and within what timescale restoration of the system must be achieved. Often the response leads to 'not really essential' and 'within 24 hours'.

           

          If you are along these lines then you would need to consider taking full backups for the Core and the DB and getting them replicated to a remote location so that they can be restored to the remote Core and Database manually.

           

          Your task is made harder by the use of a physical system. LANDESK is not really a multi-core system with DR capabilities. Multi-Core is for operational/admin reasons, not for DR reasons. Each Core is different and clients do not know about anything other than their own Core. This means that the simplest DR setup is to get 'The Same' Core running on the DR site. So we are looking more to an offline copy of the Core that can be spun up when needed. Using Virtual versions of the Core can make this sort of task easier. The DB is less important as it can be restored anywhere as long as when the Core is brought up it gets re-pointed.

           

          Much of what LANDESK holds is not affected by a period of data-loss. If we lose some scan data, clients will recover by re-sending. If the Core has lost some historical info for tasks then clients will often cover that with what they hold in their local DB to stop re-installation.. Not perfect but only occasionally would this mean a negative impact worse than sending an out-of-sync inventory scan.

           

          If you have gone through all of this consideration, fair enough, but I just wanted to put it out there just in case.

           

          Mark McGinn

          MarXtar Ltd/MarXtar Corporation

          http://landeskone.marxtar.co.uk

          LANDESK One Development Partner

           

          The One-Stop Shop for LANDESK Enhancements

          1 of 1 people found this helpful
          • 2. Re: How do you sync device inventory across 2 core servers for Disaster Recovery?
            Apprentice

            If it were up to me, we wouldn't be using physical core servers but our VMware environment isn't up to the task of hosting a server with the hardware requirements that our core server requires. It's something we are working on though!

             

            I understand the support piece as we are currently dealing with the remnants of the first attempt at using the old process. Support is doing what they can to help, but as you said, it's not something it's designed for. There are a lot of left-overs if the DR server doesn't have the same server name which can cause a lot of unforeseen issues.

             

            The functionality that is required to be up and running in a DR situation are Anti-virus, software distribution, imaging, and most importantly Remote Control. Imaging doesn't require anything outside of Provisioning which can be replicated using Core Sync. Remote Control continues to function as we use a DNS alias for our core server which takes effect as soon as we go into the DR situation. Anti-virus is set to go out to the internet if the core can't be contacted so that isn't really an issue either.

             

            The software distribution piece is where we run into issues as the amount of time it can take for the inventory to scan itself back in can take a few days. The last time I attempted the failover test it took 3 days to get the inventory scanned in as the core server ended up getting back-logged with "Full" inventory scans.

             

             

            If I can get something setup and fully functional, I can create documentation for it so anyone else running physical servers could have something to reference.

            • 3. Re: How do you sync device inventory across 2 core servers for Disaster Recovery?
              Rookie

              Sorry to ask my question here, I don't know how to post one. I am new Landesk administrator/operator and would like to know the following:

              we have just discovered that our computers appear on 2 serves that is server 52 and 51 and we would like to have them appearing in one server which is 52, how do we go about doing that?

               

              Whats happening now is in one site we have devices appearing on both servers, and this now forces me to run report from both servers.

              • 4. Re: How do you sync device inventory across 2 core servers for Disaster Recovery?
                Apprentice

                I guess I need to better understand what the end goal is for you. Are you trying to remove the second server (51) or are you trying to have 52 have a full inventory from both and 51 only has a smaller subset?

                • 5. Re: How do you sync device inventory across 2 core servers for Disaster Recovery?
                  Rookie

                  Yes we want to remove the second server 51, and have the full inventory on 52 so as to avoid managing our devices from two different servers

                  • 6. Re: How do you sync device inventory across 2 core servers for Disaster Recovery?
                    phoffmann SupportEmployee

                    Couple of points:

                     

                    • MarXtar comments regarding DR are spot on. By and large, it's mostly a non-issue, due to the clients sending up-to-date information status back. (I'll mark his answer as correct for the purposes of this thread too, since that didn't seem to get done earlier).
                    • There *ARE* ways that you can duplicate this information across -- things like storing Inventory scans & then copying them on to another Core.
                    • Generally, it's pretty unlikely that both your LD Core and your DBMS server blow up at the same time. It's about 5 minutes of effort to hook up a new Core to an existing database (a few settings need to be changed) and the rest can be helped a lot with DNS aliases.

                     

                    • For Zeus - be aware that 2 different Cores will *NOT* be able to manage a single client, unless one or more of the following is true:
                      • The client has server-certs (.0 files) from BOTH Cores.
                      • Both Cores share certs from each other.

                     

                    The certificate(s) are your biggest prize and the crucial piece to back up. If you lose your Core certs, you're in a fair bit of trouble, as the clients will refuse connections / being managed by a Core that can't authenticate with them. You can request information from support on this if need be ... I've not so far brought myself to post this publicly as a general article, because I can see a few things going bad with that.

                     

                    Registering additional certs is not a tricky affair - I just wanted to point out that there isn't a "oh, it's LANDesk - that's fine" approach to this. Security has been a fair bit tighter than that for quite a many years now .

                     

                    Hope that helps.