1. I disagree, a good use case is the UCS, where I am not swapping out a SAN but swapping out blades.

    Replacement becomes very simple with this approach, and I am not managing several independent spinning disks with separate raid etc. The OS follows along, now I know that we could potentially have to deal with driver issues, etc.

    Since you are not a fan of Hyper-V a VMware installation is fast, and really Windows isn’t terrible (assuming you have a patch management system in place).

    If the biggest issue you have on a SAN migration is installing new hypervisors, you are doing great.

  2. Anon E. Mouse,

    When using UCS, it still has to tie into the SAN. Yes, you can swap out blades all day long — just like I can take two SD cards and/or drives and swap them between blades. You are still in the same boat when you need to replace your SAN. Yes, it is not a huge deal to rebuild an ESXi server (assuming you do it right – Enterprise licenses, host profiles, vDistributed Switches, proper vMotion networking, etc.) and the same goes for Windows (though this does take more effort – or cost if you use VMM). However, what happens when you walk into a larger environment that has 16 ESXi servers and 16 Window Bare Metal (why do people still do this?) blades/servers. Standard licensing — not Enterprise. Now, you are rebuilding 32 nodes (or some form of mirror/replication setup) to replace your SAN. This adds up quickly – very quickly in terms of cost and/or time – which is again, cost.

    P.S: If you are going to do something, do it right. If not, call me. I will do it right for you. 😉

    P.P.S: Hyper-V has its place – it is called Dev/Test/UAT/QA – and it works well in that place. Hell, you can even make it work in production — have at it. You can even call me when it breaks. I know how to fix it. 😉 …and when you are ready to replace it with VMware, you can call me too.

    On a side note, I do allow anonymous comments, but really — we are all friends here, so feel free to share. 😀

Leave a reply canel