• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

VMware vs KVM (CentOS/RedHat/Ubuntu)

parityboy

Limp Gawd
Joined
Nov 13, 2010
Messages
390
I'm researching virtualisation options for an SME. I know there are a lot of options out there, but I've narrowed the choices down to these two.

Has anyone here had experience of either these two, specifically in a production environment? What's your impression of them?

Many thanks. :)
 
Stick with Vmware. Not just because of being VCP but because of the less headaches you will have.
Linux is great and all but its hard to find good support and has more dependencies. Unless you have a good support for Linux rollout and or you have talent that can support this I would stick with a proven commercial solution.
 
You can try Proxmox VE or RHEV - both have paid support subscriptions. With RHEV you pay per managed socket, with no limitations to cores or RAM. A lot cheaper than Vmware.
 
You have to look at the requirements first, there are plenty of ways to virtualize but how much do you need outside of that?

I'm obviously a VMware evangelist, but there is an impressive stack of products once you step away from the hypervisor itself.
 
OK, the story is this. I'm setting up a software development company. Obviously we'll start small, but I'm looking at getting the infrastructure as right as possible so that as we grow, the disruption will be minimal.

This is why I've taken the decision to virtualise from day one. To me this makes sense, so that as the business grows we can accommodate more and more-demanding requirements by having the flexibility to do so.

To this end, I'm thinking of an infrastructure with the following VMs:

- Directory Server. This will run DNS and LDAP services, and possibly Kerberos.
- File & Print Server. This will run NFS & CIFS and will hold the user directories. It will also run a CUPS server.
- Comm Server. This will run email and IM, and possibly VoIP services (TeamSpeak, Asterisk).
- git server. This will hold the source code and other IP.
- Backup Server (possibly). This will run Zmanda.

I'm looking at using Ubuntu Server with KVM basically to save a few coins - VMware is very nice, but it's expensive, so I'm looking at using Convirture OSE, to manage the hosts. I found this product last night. :D

Hardware-wise I'm looking at either two Dell PowerEdge R510 servers (one each for VMs and SAN - the SAN box will be running either NexentaStor or QuantaStor) or one Dell PowerEdge R510 and one Dell PowerVault MD3200i, with Dell PowerConnect GigE networking infrastructure connecting it all up.

My intention is to have the VMs stored locally on the VM server on a RAID10 array, rather than on the SAN box. However, the SAN box will present iSCSI targets to be mounted inside each of the VMs, so for example the File & Print VM will mount the iSCSI target for the user directories, which in turn will be a volume on the SAN.

Additionally the SAN box will be used to store backups of data from the VMs, as well as backups of the VMs themselves. This is why I want to use the iSCSI initiator-inside-the-VM approach. It means I can keep the VMs at 32-64GB each, while keeping the data they manage outside of the VMs.

As far as commercial support goes, I think I'm in a good position. All of the hardware will be Dell, Ubuntu Server is certified on Dell hardware, and ConVirt is certified on Ubuntu Server, so there are as few people as possible to shout at. :)
 
My intention is to have the VMs stored locally on the VM server on a RAID10 array, rather than on the SAN box. However, the SAN box will present iSCSI targets to be mounted inside each of the VMs, so for example the File & Print VM will mount the iSCSI target for the user directories, which in turn will be a volume on the SAN.
This is convoluted and unnecessary. Why have the user data plus the client filesystem on the SAN? Use a file-centric approach instead, i.e. NFS or CIFS.
 
This is convoluted and unnecessary. Why have the user data plus the client filesystem on the SAN? Use a file-centric approach instead, i.e. NFS or CIFS.
Also along these lines keep all your data stored off the hypervisor host, except the hypervisor OS. This way you can keep stacking hosts into the environment as you grow and the guests don’t care where they run. This can allow you to scale faster.
 
Back
Top