Jump to content

Recommended Posts

Posted

This has been going round in my head for a while because i've been a little unsure about it all and only now i've managed to get it out on what we have and what is the best way to move forward with this.

 

Our current HyperV setup consists of 2x Virtual Hosts and 1x SAN which are starting to get a bit long in the tooth (we are currently sat at 7 years and coming into year 8 at summer 2022) and we are continuously paying for expensive warranties which are going up year on year and the parts for the servers and the host are starting to get harder to source.

 

The HyperV hosts are a pair of HP DL380 G8 each with 2x E5-2620 v2 Processors | 128Gb Ram | 2x Raid1 Local drives at 300gb each running Windows Server 2012 R2. I would love to do an in-place upgrade to 2019 but if i did im not sure what would go wrong, if anything at all so i feel we are stuck on 2012r2 and as i didn't set these servers up i'm unsure how they have been configured. I'm mentioning about doing an in-place upgrade because i'm having issues installing windows updates on the Hosts now.

 

Both these servers are connected to our DotHill R/Evo 3000 SAN via iSCSI with 12x 1Tb drives so giving us about 9Tb of usable space which has been no trouble at all. The SAN is connected to a pair of SAN Switches via 4x 1Gb ports on the SAN and to be fair, it's never missed at beat unless the power has failed. We have had a little trouble with one of the HyperV hosts though - there is also a QuorumDisk on one of the hosts.

 

The only gripe we have is when we start working on the VMs to do windows updates or the VMs haven't been used for a while or even schedule reboots the VMs take sooooo long to respond or install the updates. Also, when we do reboot the VMs, not all the services start so we have to painstakingly wait for each server to finish it's reboot, then wait for the services screen to appear and work through the services which should have started such as the Sophos AV ones, the DFS Replication or even the VEEAM services. It's like we can only work on 2 or 3 at a time. I mean, it took nearly 3 days to perform windows updates on our 17 VMs over the xmas holiday. When we work on the physical hosts, they are as quick as anything. The only thing i can think is the issue is the SAN drive RPM speed, i seem to think they are only 10,000rpm which could be having an effect.

 

We did have some teething issues when the kit got installed and i wasn't 100% confident with the install but we let it be and now i'm wondering if this is the result.

 

So moving forward, im not sure what the best way to go is, what have others done recently or what are others planning on doing in the future? Thinking out loud, maybe an assessment is required by the supplier who we still buy the warranties through might be the best course of action.

Posted (edited)

we have a performance tiered HP SAN (SSD.15k>10k >10k MDM) which is transparent to the cluster.

this serves six Hyper-V hosts running Server 2016 Data centre (HP DL360 Gen 9's (x4 & 2 x Gen 8's). our hosts were added in pairs and have from196Gb up to 300Gb of RAM

all connect via 10Gb Copper ethernet to two multi-way 10Gb panels to the SAN which also has 2 x controllers at 2 x 10gb each.

 

I'm planning on keeping this setup but replacing the older hosts with faster newer HP DL360's next year but with more cores and multiple CPU's and LOTS more RAM

the new Gen 10 DL360's are much more capable too!

You could look at creating a second cluster based on two new machines and migrate your VM's across?

 

Steve

Edited by Ex-MGSTech
Posted

Look at a S2D setup (Storage Spaces Direct). Go with SSD's for cache then HDD's if you have a lot of storage.

 

Hyper-V then works pretty much as it does now for you, just working off the shared storage space generated between your two servers.

 

Put a DC locally hosted on each host server just in case the Shared Storage needs to contact the DC before starting. You likely only need two DC's anyway.

Posted (edited)
Has anyone got any ball park costings for getting a 3 node 40Tb S2D / Hyper-V setup stood up?

 

£80-90k

 

3 x Lenovo SR630 With 2 overkill Xeon gold chips and 512Gb ram

1 x Lenovo DE4000F all flash array - 56TB (Lenovos DE series are a collaboration with Netapp)

Badass switches with 24 ports of SFP+ (plus the modules)

5 year warranty.

Installation/configuration

 

comparable dell/HPE were almost double the cost and didn't consist of a full flash array.

 

This was almost 2 years ago now and I've not had a single issue. They're rapid.

 

Re used the old SAN as a NAS, old hosts rebuilt as standalone for testing etc.

 

I can share the details of the company I went through if you're interested via PM.

Edited by slugshead
Posted
Would go for storage spaces , I’ve always been HP for servers but dell have impressed me recently with there r640 great for hyper v / VMware

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...