Announcement

Collapse
No announcement yet.

Server Redundancy/Failover/Replication

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Server Redundancy/Failover/Replication

    Hi All

    At present we have 1 Server running SBS 2011, 1 x Server 2012R2 running Sage 200, 1 x Server 2008 running bespoke software, we want to upgrade the system.

    We want to build in Redundancy/Failover/Replication/Hyper V/DFS in the event of hardware/software failure etc

    We want to upgrade to: 1 x Server 2012R2 DHCP, 1 x Server 2012R2 with Exchange 2013, 2 x Servers 2012 R2 (Sage & Bespoke)
    Do we double the purchase above and then install service/roles etc or use Hyper V Replicate, any suggestions would be appreciated, cost really is not a major issue, they just want to make sure that if anything goes down it won’t effect the users

    Thanks in advance

    Andrew

  • #2
    If you can afford it, consider HyperV replicas, so 2 beefy physical servers with all 4 VMs on each (2 active, 2 replica copies)

    At present you don't have any redundancy in roles, so redundancy in VMs is your best option
    Tom Jones
    MCT, MCSE (2000:Security & 2003), MCSA:Security & Messaging, MCDBA, MCDST, MCITP(EA, EMA, SA, EDA, ES, CS), MCTS, MCP, Sec+
    PhD, MSc, FIAP, MIITT
    IT Trainer / Consultant
    Ossian Ltd
    Scotland

    ** Remember to give credit where credit is due and leave reputation points where appropriate **

    Comment


    • #3
      What would be a recommendation for a beefy server, was also thinking of putting exchange into the cloud, which could make all of this a bit easier, redundancy is a must, there are about 30-32 users so i assume i am going down the right path?

      Comment


      • #4
        Hello,

        I agree with the Ossian

        Originally posted by Ossian View Post
        If you can afford it, consider HyperV replicas, so 2 beefy physical servers with all 4 VMs on each (2 active, 2 replica copies)

        At present you don't have any redundancy in roles, so redundancy in VMs is your best option

        Maybe it's a good solution to go in cloud with Exchange . But what about the other Servers, Softwares? HYPERV-Replica will be give you a complete solution in your problem.
        Last edited by ktzouv; 23rd November 2015, 08:29.

        Comment


        • #5
          ktzouv Cloudy Exchange depends very much on who you trust with your data, and your legal requirements surrounding it. Personally I prefer to make the TLAs work a bit to get hold of my information

          Netbase Look into capacity planning tools as the number of users is only one of the factors involved.
          As a quick and dirty solution, look at around 6-8 processor cores and 32-64 GB of RAM, at least 2 NiCs and storage as appropriate for current and future needs
          Tom Jones
          MCT, MCSE (2000:Security & 2003), MCSA:Security & Messaging, MCDBA, MCDST, MCITP(EA, EMA, SA, EDA, ES, CS), MCTS, MCP, Sec+
          PhD, MSc, FIAP, MIITT
          IT Trainer / Consultant
          Ossian Ltd
          Scotland

          ** Remember to give credit where credit is due and leave reputation points where appropriate **

          Comment


          • #6
            So would the 2 Servers be standaone, then on 1 of them in VM the DC, and other member servers, then on the other 1 the same?

            Comment


            • #7
              As long as you are using 2012R2 (maybe just 2012) there are no issues with a physical host being a domain member with a virtual DC on it), but basically yes, all 3 or 4 VMs on each server (mixing replicas and live on both to load balance)
              Tom Jones
              MCT, MCSE (2000:Security & 2003), MCSA:Security & Messaging, MCDBA, MCDST, MCITP(EA, EMA, SA, EDA, ES, CS), MCTS, MCP, Sec+
              PhD, MSc, FIAP, MIITT
              IT Trainer / Consultant
              Ossian Ltd
              Scotland

              ** Remember to give credit where credit is due and leave reputation points where appropriate **

              Comment


              • #8
                So if I was to set this up in a lab test first, The physical host has to be Server 2012 R2 or Windows 10 ?, then with the Hyper V Manager on it and all the VMs

                Comment


                • #9
                  If you're doing it in a test-lab, then Windows 10 might be ok. However, I suspect it's going to be missing many of the features you'd want in your live environment. I'd suggest just sticking with 2012R2 in your lab.
                  So you're looking at 2xbig chunky servers, like alreasdy suggested, 64GB of ram. Shared disk backend, so an iscsi target or similar.
                  Windows 2012 R2 Enterprise Edition (2 licences) installed on each server. Configured exactly the same, with the same vnetwork names etc. Ideally, a shared disk backend (someone might tell me where this is wrong) with clustering setup
                  Then install your guests as needed. set up two to be primary instance on each server, and ensure they replica to the other.
                  This way, you've got SVR1 and SVR2 primary on HV01 and replicad to HV02. On HV02, you have primary of SVR3 and SVR4 and replicas of svr1, svr2.
                  You'll also need client access licences for Exchange, and for Server 2012 R2.
                  Each HyperV host is likely to need minimum 4 network interfaces - 1 for management, 1 for Vguests, 1 for iscsi, 1 for spare incase you need separate vlans etc.

                  for real, true redundancy you'll also need:
                  Matched, clustered switches with relevant network teams
                  So, Switch1 has a port going to each nic on HV01 and each nic on HV02. Switch2 has the same, and the NICs are configured in teamed active/passive.
                  In this way - if you lose one switch, you don't lose it all. If you lose one NIC, you don't lose it all.

                  Redundant or paired firewalls in some way for internet access. I wouldn't even begin to know how to configure this.. and redundant internet feeds also.


                  Obviously also redundant power supplies and redundant power feeds. (separate circuit breakers at a minimum, ideally separate distribution boards and even better, separate feeds from the grid.)


                  don't forget to include your backup platform (replication is not backup; redundancy is not backup) - perhaps daily Guest snapshots and 15 minute data-snapshots.

                  If your disks are in-host, make sure your disks are running on real RAID adaptors with battery backed cache. Not B120 or S300s etc. SAS disks, not SATA. SSDs would be even better and you said money isn't an object


                  You will need to determine what IOPS you need for each actual vGuest and how that correlates to the IOPS the server can provide....


                  On a side note, I think I just saw that error you were talking about.. It might be related to autosave of the post. Something popped up in the middle of the screen while i was typing, but just as I hit spacebar, so it went away and I didn't see it.
                  Please do show your appreciation to those who assist you by leaving Rep Point https://www.petri.com/forums/core/im.../icon_beer.gif

                  Comment

                  Working...
                  X