VMware

Server Upgrade Advice!

Hi everyone!

Not sure if this is the correct post to ask this question, however, let me know if that’s the case.

I manage 10 servers at my work. We use them for many different stuff, but basically, they are running 50 VMs, with Esxi 6.7. The virtual machines are generally Windows 10, Windows Server and a couple of Linux. Averaging, the VMs have 3GBs of RAM, and 2 CPUs allocated. At what we need, they run at an 80% capacity with this specs.

Now, here is the tricky part: The hosts are Dell PowerEdge 410, with 64gb of RAM and 2 Intel(R) Xeon(R) CPU E5620 @ 2.40GHz which are almost 10 years old.

What would be a good replacement for these servers? Let’s say if they want to reduce the number of Hosts to 2, instead of having 10 (reduce space, heat and power consumption), having some performance speed improvement without having to sell the whole company just to upgrade the servers 🙂

Thanks for your time in advance!


View Reddit by TheCamunasView Source

Related Articles

10 Comments

  1. I’d look at 2 *identically* configured servers with 10gbE NICs, massive amounts of RAM and as much solid state storage as you can afford. (How about a RAID 5 of SSDs in each host). Arguably, if later on the budget allowed for (or business growth necessitated) a third server you could get into S2D with 3 properly designed servers and have a nice highly-available array of servers, using each other’s internal storage without the need for a SAN. (There are people here smarter than me who can elaborate). In any case I think you’ll find that 2 servers can easily replace 10 410s.

  2. I personally would look at the FX modular chassis.
    You didn’t state what your storage is and if it’s staying or if you’re getting something new.

    Call your VAR up and get liveoptics running.

    Keep in mind that if you go down to only two hosts you probably can’t survive any sort of host failure. You should design for n+1, and aim for n+2 so your regular host patching and updates don’t leave you unprotected.

  3. go for a cluster of 3 identical hosts for production and a 4th one for testing/swapping with production ones for maintenance/failures (to do your job mostly during work hours)

    check various brands for pricing since it will vary

    you haven’t said how your storage is …

  4. Honestly, depending on what your growth forecast looks like … almost any new hardware will house what you’ve got just fine.

    With this small of an environment you could probably get away with 3 hosts and one SAN. Probably $30k or so investment. DEPENDING on the needs of those Win10 VMs. VDI/VM desktops are a whole ‘nother beast.

  5. You could easily replace those 10 servers with 3. Heck, you could probably replace them with 1 but that’s not a great HA option.

    You’ve got 80 cores essentially and 640 GB of RAM.

    Something like this (doesn’t have to be Cisco, just looking at the specs)

    https://www.compsource.com/buy/UCSSPC220M5A1/Cisco-91/?src=F

    That is 28 cores per server, and 384 GB of RAM. 3 of them gives you 84 cores and almost a terabyte of RAM. You’ve got lots of storage options, but obviously flash-based memory is going to be attractive compared to the 15K drives.

  6. Also, internet connectivity will not be a problem, right? All the servers are attached to one network switch, just with one cable, the other ethernet adapter is unplugged. The iDrac is also connected to the same switch, and then, the switch is connected to the router directly (to a 500 Mbps internet line). We don’t have any connectivity issues right now, and averagely each virtual machine consumes 80mb of bandwidth (internet) per hour, or around 1 MB per second.

    Sorry for all the silly questions, I didn’t do the setup that is already made, and trying to improve things here.

Leave a Reply

Your email address will not be published. Required fields are marked *

Close