• 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.

First ESXi Production Host help

E

enuro12

Gawd
Joined
Jan 25, 2004
Messages
781
First off i'd like to thank the community for your help. I've found [H] nets much better guidance than the VMWare forums. Lots of good guys there too. Guess i just love the [H] community! :) So thanks!

[Long Explanation of the project]
So I had my big meeting today with a potential new customer. They notified me i'm in the top spot for the new contract. This is the gear I based my quote off of.

My hardware budget is roughly $10,000 for two hosts. I plan to come up with a way to do some sort of live cloning of the virtual machines on these. Each host will have a terminal server, and SQL server. I want to clone site 1 to 2 and vice versa. So if site 1 goes down we can just redirect the TS clients to site 2. I'll also be adding several layers of regular backups, but for availability that's what i'd like to accomplish.

It will support 40 users at each location, and the SQL database is around 8gb. I do not know how intensive the SQL write is, however it cannot be more than 40mb a day added.
[/end Long Explanation of the Project]

I also realize this is an overbuilt TS Server. It may get a 3rd TS server added later.

I still need to come up with a RAID 10 controller for the SSD's. I'm going to use an IBM M1015 for my lab setup. Maybe that's also good enough for this build?

1x LITE-ON DVD Burner - Bulk Black SATA Model iHAS124-04 - OEM
1x Antec EarthWatts EA750 750W Continuous Power ATX12V version 2.3 SLI Certified
1x TYAN S8812WGM3NR Quad Socket G34 AMD SR5690 MEB Quad AMD 8-Core/12-Core
16x Kingston 8GB (2 x 4GB) 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10600) ECC Unbuffered
4x OCZ Vertex 3 VTX3-25SAT3-120G 2.5" 120GB SATA III MLC Internal Solid State Drive
4x AMD Opteron 6128 Magny-Cours 2.0GHz Socket G34 115W 8-Core Server Processor
 
My suggestion is to use two hosts and a shared NFS NAS on each side. Then do like Essentials licensing so you can do things like HA. One big box is good...until it fails.
 
When you say site do you mean different physical locations or the ESXi host?
 
NetJunkie said:
My suggestion is to use two hosts and a shared NFS NAS on each side. Then do like Essentials licensing so you can do things like HA. One big box is good...until it fails.
Click to expand...

I think he says he is going to have two hosts total.

Sounds like he wants FT between the two locations.
 
It will be two hosts in separate locations. Connected across a VPN on 10mb DSL

I really don't think a NAS is going to be required for 40gb of data.
 
How are you going to replicate them? Same layer 2 network in both places or will you need to deal with IPs on a recovery?

Personally..I don't want a bad DIMM in a server to cause me to fail to a second site with manual work. That's why I say two hosts in the primary site...and one host in the secondary for DR failover.
 
enuro12 said:
It will be two hosts in separate locations. Connected across a VPN on 10mb DSL

I really don't think a NAS is going to be required for 40gb of data.
Click to expand...

This is beyond my knowledge but maybe look into the regular MS clustering and HA features for this setup?
 
NetJunkie said:
How are you going to replicate them? Same layer 2 network in both places or will you need to deal with IPs on a recovery?

Personally..I don't want a bad DIMM in a server to cause me to fail to a second site with manual work. That's why I say two hosts in the primary site...and one host in the secondary for DR failover.
Click to expand...

They are on the same layer 2 network.


If 1 host went down, and the other host could pickup the TS sessions by simply pointing to a new IP from the clients while i rebuild the host, that would work fine. I just like the idea of not having to 'restore' anything to be back up and runing while i'm rebuilding the host.

The only data changing is SQL for the most part. So maybe SQL replication and nightly rsync for documents is as good as anything.
 
Last edited:
Can do something like Veeam Backup/Replication to mirror the VMs to the other site. Works well. Since you are extending the L2 network it would just be a simple boot of the VM on the other.

If the primary site is totally down you need to look at routing, DNS, IP connectivity, etc. But getting the VMs up over there is easy.
 
Connectivity doesn't appear to be an issue. Our firewall manages DNS, DHCP ect.

Veeam appears to be exactly the ticket i was looking for. Thanks!


Does anyone have any comments on the hardware setup?
 
Only problem with Veeam is that the failover is not automated. And also if you do failover to your "Other/DR Site", you cannot simply fallback to your main site, you will need to resync the ENTIRE VM's back to the main site as opposed to just the changes.

With all that said for your requirements, Veeam is great. It's about ~600 per proc.
 
No site failover should be automatic. Site failover is not HA...that's almost always a DR event that should be manually initiated.
 
Sorry if I misunderstood your requirements but if what you need is HA for the SQL server databases you could also use one of the SQL server HA solutions.

One option: you can set up mirroring (async if you want to keep the traffic down) or (sync with a witness server if you need fast automatic failover)
 
Firstly I would not build these as whiteboxes. Two Dell R310s (or R415s if you want AMD) could do everything you need, and fit your budget just fine. The reason I do not suggest whiteboxing in a prod environment is that VMware offers very limited support for whitebox hardware. For a home lab it is fine, but I would never put it into production. Also just remember that for a TS / SQL server RAM is going to be far more important then CPU. I would probably go with whatever Quad core hyperthreaded (or equivalent AMD) was cheapest, and spend the rest of my budget on 24 or 48GB of RAM. Then for a 40 user TS I would chop at the very least 16GB of RAM to it, and then throw the rest at the SQL server.

Secondly if you are looking to do hot site cold site there are a couple ways you could do that.
-First (and the way I would do it) would be to have both server up and running 100% of the time. Use a session broker server to LB between your terminal servers. That way there is automatic failover for your TS servers and your users just reconnect instead of you having to do a bunch of work to get the second server up and running. For the SQL again do a cluster. One side goes down, not a huge deal.
-Secondly (the way I would do it if #1 got to complicated) use vCenter server to automatically migrate VMs to your second site. Now to do this it required some fancy storage and some licensing, however HP did a demo a while back showing proof of concept and it was pretty awesome to watch one site go down, and the other one come up within 30 seconds.
-Third (sorta #2 with less cash involved) I would use Veeam to basically rsync my datastores and if site 1 goes down, I could manually bring site 2 up. I wouldn't prefer this because I don't like getting up in the middle of the night to fail servers over, I like finding out why they failed and fixing them when I get in.
 
C7J0yc3 said:
Firstly I would not build these as whiteboxes. Two Dell R310s (or R415s if you want AMD) could do everything you need, and fit your budget just fine. The reason I do not suggest whiteboxing in a prod environment is that VMware offers very limited support for whitebox hardware. For a home lab it is fine, but I would never put it into production. Also just remember that for a TS / SQL server RAM is going to be far more important then CPU. I would probably go with whatever Quad core hyperthreaded (or equivalent AMD) was cheapest, and spend the rest of my budget on 24 or 48GB of RAM. Then for a 40 user TS I would chop at the very least 16GB of RAM to it, and then throw the rest at the SQL server.
Click to expand...

I cant help but wonder if you bothered looking at the hardware i plan to use. The board i'm going with has also been certified for Vmware. Although if I go your route for only $274 more, i can have 28 fewer cores, and 96gb less ram.
 
Neutrino said:
Sorry if I misunderstood your requirements but if what you need is HA for the SQL server databases you could also use one of the SQL server HA solutions.

One option: you can set up mirroring (async if you want to keep the traffic down) or (sync with a witness server if you need fast automatic failover)
Click to expand...

Thanks for the reply. After going over all the specifications of the current network setup i've decided the only real data i need real time copies of is SQL. So i'll be doing some mirroring and rsync for my file structure. The amount of data being changed in SQL is rather small (Less than 100mb a day). So I think this will work excellent for the customer in question.



I've really appreciated all the input guys. I wish some if you lived near me, it'd be nice to sit down and have a intelligent conversation regarding technology...
 
You must log in or register to reply here.
Back
Top