vSAN1

 View Only
  • 1.  Is it possible to rebuild a failed node in 3 node VSAN ?

    Posted Feb 14, 2019 12:02 PM

    Hi,

    just wondering if you can replace a failed node in 3 node VSAN,

    Nutanix clearly states you can't replace dynamically a failed node in 3 node cluster,  Nutanix Portal

    So I was wondering if you could do that with VSAN ?

    Thanks



  • 2.  RE: Is it possible to rebuild a failed node in 3 node VSAN ?
    Best Answer

    Posted Feb 14, 2019 01:26 PM

    Hello fnature

    Yes, of course this is possible and supported in vSAN - either adding a replacement server (and creating new Disk-Groups) or replacing failed components on the 3rd node would suffice.

    Bob



  • 3.  RE: Is it possible to rebuild a failed node in 3 node VSAN ?

    Posted Feb 14, 2019 02:20 PM

    thanks Bob makes sense



  • 4.  RE: Is it possible to rebuild a failed node in 3 node VSAN ?

    Posted Feb 21, 2019 08:01 AM

    In Nutanix you can’t go from a 3 node cluster to a 2 node cluster but of course you can repair and replace the failed node in a 3 node cluster, without downtime.

    If you had a 4 node-cluster you could remove the failed node and continue running the cluster with full functionality as a 3-node cluster.

    That is what the document on the Nutanix portal is about.

    What important is that in case of a node-failure in a Nutanix 3-node cluster, the data will be rebuilt and once rebuilt the cluster can sustain additional disk-failures without downtime/dataloss and node-failure without dataloss.

    Disclaimer: Nutanix employee but try to stay objective.

    // Linjo



  • 5.  RE: Is it possible to rebuild a failed node in 3 node VSAN ?

    Posted Feb 25, 2019 05:04 PM

    fnature just as TheBobkin states, you can absolutely replace a failed 3rd node in a 3 node vSAN Cluster.

    If you only need to rebuild the 3rd node, when you reinstall ESXi, you'll see an asterisk beside each of the currently consumed/used for vSAN disks.

    When the host is reinstalled, and all networking has been restored (vSAN VMkernel interface configured), the disk groups will be seen again.

    If you replace the failed node with a net new (not used in the cluster before) node, once ESXi is installed, added to the cluster, and networking (vSAN VMkernel interface) added, you'll need to create any new disk groups.

    When either of the above are performed, I'd also suggest invoking a "Repair Objects Immediately" operation from the vSAN Health Check UI.

    Can't really speak to Nutanix, as this is a VMware community.