2 Replies Latest reply on Dec 13, 2019 4:52 AM by pragg12

    ESXi 6.0 host - Some VMs go offline intermittently

    cschneer Lurker

      Recently, like some of you, I've been going through the process of upgrading/migrating my services from Server 2008 to Server 2012-->2016. I also recently upgraded our ESX host to version 6.0, and will upgrade again to 6.5 (the newest version my server is compatible with) in the coming weeks when I can schedule a maintenance period.

       

      Some of my VMs have now been going offline intermittently. I have to disable and reenable the NIC on the VM to bring it back online. I've been checking through windows logs but cannot find much. I've heard recommendations to upgrade VMware tools, which I've done for these servers.

       

      What can I do to trouble shoot this?

       

      Are there logs on the host that will give me data on this?

       

      Thank you all.

        • 1. Re: ESXi 6.0 host - Some VMs go offline intermittently
          Amin Masoudifard Expert

          ESXi hosts and VMs with Windows Server guest OS have completely two different upgrading methods ... When you upgraded the host, did you upgrade the hardware compatibility of those VMs or not? (before upgrading the VMware tools)

          Instead of windows log, please check the vmware.log file in the VM directory path inside the datastore, especially time duration of upgrade actions

          Please mark my comment as the Correct Answer if this solution resolved your problem
          • 2. Re: ESXi 6.0 host - Some VMs go offline intermittently
            pragg12 Hot Shot
            vExpert

            Hi,

             

            Why are you not going for direct upgrade of ESXi to 6.5 instead of current version -> 6.0 -> 6.5 ?

            Since you are ESXi upgrade in latter way, you will have to upgrade VM hardware version upgrade in same way as well. Like this: Current version -> 11 -> 13. So double downtime per VM.

            Also, what NIC are the VMs using ? VMXNet should be there instead of E1000 unless there's a particular reason.

            What are your observations on the VMs which are facing the issue ? Any common pattern you observed; like same OS, or NIC or host or cluster ?