DX Unified Infrastructure Management

 View Only

Nimsoft questions arroun distributed architecture with HA - Windows Cluster/VMotion/native

  • 1.  Nimsoft questions arroun distributed architecture with HA - Windows Cluster/VMotion/native

    Posted Feb 07, 2013 03:28 PM



    I'm Steffen, new to the Nimsoft Monitor product and this is my first thread in this forum as well.


    For my 1st Nimsoft project Im looking for the right and most easiest overall design. After reading all about Nimsoft architecture and HA documents, I was able to get, it leave me a lot of questions how the overall distributed while completely redundant architecture should look like.


    Distrubuted Design consists on this components and requirements:

    -          UMP/UR

    -          Central PriHub

    -          Several RemoteHubs (Collectors) at subsidiaries

    -          No single point of failure, totally all Nimsoft components has to be redundant

    -          SSL-Tunnels between central site and remote subsidiaries

    -          Northbound VIP for UMP-URL, Southbound VIP for traps and syslogs without can use external Loadbalancers


    HA in case of Windows Cluster (“” are quotes from install guide):

    1. “Create Virtual IP address assigned to the virtual Nimsoft service” => Is this different to the Cluster-VIP at Windows-OS level (node)? !!!”When you specify the network interface, enter the virtual IP address you assigned to the virtual Nimsoft service. Do not use the virtual IP address of the node or its physical IP address. “!!!
    2. Shared disk/iSCSI target (typically SAN, NAS or RAID array) configured => Do we need this really also in case of having an external DB with access to the DB-cluster-IP? I learned the NIS is the only shared resource for Nimsoft.”
    3. “Recommended: Do not install any Nimsoft consoles (primarily Infrastructure Manager) on the cluster nodes. Install these on a separate workstation.” => Why?
    4. Are SSL-tunnels between central and remote clustered hubs supported? Could the SSL-Tunnels be bound to the VIP (node or service address, if different), saving Backup-tunnels?
    5. Clould also the UMP and the remote SiteHubs be implemented as Windows-Clusters?

    HA in case of VMWare ESX VMotion:

    1. Could all components be set up as would them be standalone?
    2. Could VMotion at Primary Hub (Central Site) be connected via SSL-tunnels with one standalone Cluster at one site (RemoteHubs) without existing VMWare-Infrastructure?


    HA with Ha-probe:

    1. Is the use of ha-probe limited to the Primary-Hub at central site only?
    2. What kind of HA has to be used for the remote collectors and UMP server pairs?
    3. Do we need SSL-tunnels with backups cross wise? Do they need different ports away from standard 48003?

    -          PriHub-primary <=> SiteHub1-primary

    -          PriHub-standby <=> SiteHub1-primary

    -          PriHub-primary <=> SiteHub1-standby

    -          PriHub-standby <=> SiteHub1-standby

    4. In one Nimsoft Document, called Ha-probe.ppt, is at Step 7 outlined: “Drop HA probe on secondary”. What could this mean? I’snt this a dissent to ha-probe should run on Secondary?



    thx for answers in advance,