Building a Lab Part 3 Configuring VMWare ESXI and TrueNAS Core

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

pinkanese

New Member
Jun 19, 2014
27
10
3
33
Am I missing something, or is the section about ZFS absent from the article?

"Before continuing the configuration, we must first talk about ZFS"
 

Callan05

New Member
Nov 8, 2018
18
7
3
I find this series a bit confusing, to be honest.

There was very little statement of requirements (how many VM's, how much storage, ram/CPU, availability, usage), then we have started to build a cluster with a single storage node as a single point of failure.
At this point, true nas updates are going to take offline any servers needing that iscsi storage.

Was there any reason you didn't go down the path of hyper-converged storage, with something like star wind?
This could have reduced the footprint and power consumption, potentially also the complexity.

Maybe you are still getting to this, but I would expect readers should be thinking about their requirements upfront before buying hardware and starting to build their own virtual environment.

A part 0 might be useful once you've finished the series?
 

EngineerNate

Member
Jun 3, 2017
68
16
8
35
If it's specifically built as a lab I'd think concerns about storage availability would be slightly less critical.
 

nickf1227

Active Member
Sep 23, 2015
197
128
43
33
I understand and appreciate the concerns.

On ZFS:
This article, in its entirety was written to almost 10,000 words. Patrick and I decided to spin the primer on ZFS to a dedicated piece not part of the lab. That is already written and edited and will be coming soon. I had actually wanted that piece to be released first, but there was a long delay between part 2 and 3 and Patrick wanted to make sure this got released.

On requirements:
I specifically stated in part 1 that this was for a small office scenario with 2-3 VMs, or for lab use.

As for the storage server being a single point of failure, and hyperconverged, downtime etc. I understand and appreciate the criticism there. I am writing a parallel 2-parter doing everything with in a 3 node Proxmox cluster with ZFS/Gluster. I've done alot of the work here already and it should be coming soon.

I am trying to show the various alternative scenarios and models you can use to provide these types of systems. These are not things I hadn't thought about. This series represents the "traditional" model where you have external storage connected to a hypervisor.

After that the intention is to write about Kubernetes, the WSL2 and to provide a cross over between this type of content and project mini micro. There are many ways to do these types of things. There are pros and cons to alot of them. I'm try to illustrate that fact as best I can. I am also trying to reach a wide audience here. Alot of members know alot of what I have talked about, but alot of newcomers don't. I'm trying to appeal to both audiences, but I can never make everyone happy.

I can only say so much 3,000 words at a time :)
 

Callan05

New Member
Nov 8, 2018
18
7
3
On requirements:
I specifically stated in part 1 that this was for a small office scenario with 2-3 VMs, or for lab use.

As for the storage server being a single point of failure, and hyperconverged, downtime etc. I understand and appreciate the criticism there. I am writing a parallel 2-parter doing everything with in a 3 node Proxmox cluster with ZFS/Gluster. I've done alot of the work here already and it should be coming soon.

I am trying to show the various alternative scenarios and models you can use to provide these types of systems. These are not things I hadn't thought about. This series represents the "traditional" model where you have external storage connected to a hypervisor.

After that the intention is to write about Kubernetes, the WSL2 and to provide a cross over between this type of content and project mini micro. There are many ways to do these types of things. There are pros and cons to alot of them. I'm try to illustrate that fact as best I can. I am also trying to reach a wide audience here. Alot of members know alot of what I have talked about, but alot of newcomers don't. I'm trying to appeal to both audiences, but I can never make everyone happy.

I can only say so much 3,000 words at a time :)
Thanks Nick,
I think if the purpose was : here's a hands on experience where you can learn about esx cluster with shared storage, then great.

I was worried somebody might copy this for their small office setup, and get into trouble later on.

In article 1, you mentioned ad and DNS as the use cases (+ other), if you duplicate these 2 on each host, running on local storage, many of my availability concerns go away. ('core' services have resiliency with no SPOF)
That leaves only the 2-3 VM's left

Sorry if I came across a bit blunt earlier, just trying to help anybody who might come across this for use in small or home office, as opposed to education/learning.
 
  • Like
Reactions: nickf1227

nickf1227

Active Member
Sep 23, 2015
197
128
43
33
I understand man, truly.
Here's what I think. Alot of people start off their IT careers by building PCs for friends, doing basic helpdesk roles, etc. Alot of them try to break out and do more complex stuff and try to do consulting work for small businesses to build up their resumes. Alot of small businesses don't have IT staff, don't know anything about IT and just need some occasional help.

If a small company of 4 or 5 people reached out to me to build a system for them to provide a specific application to their office, this lab represents something like what I would do, plus backups. Let's imagine that the client is running everything off their local PC and the rest of the folks in the office are accessing a share on that desktop, or they have a single drive bay consumer NAS.

They want to hire you to modernize their work flow, and maybe integrate a new line of business application. They have no idea that what they have been doing thus far is crazy. Thye don't understand. Some places I've consulted for had no backups...except a small fraction of their data was stored in paper form.
If you want the job and you try to tell them you can't get them off the ground without spending money on 3 separate servers for an HA cluster, plus networking/firewalling, plus backup, plus wiring, etc....they aren't always going to hire you. If you instead give them a solution to get them started with a path to move forward as the business needs grow, you have created a consistent customer, rather than a 1-off one.

While I certainly agree with you that in most cases I wouldn't want to build something without considering availability. That's just not always an option for these small businesses or for people getting started building a lab in their house while they work part-time jobs.

IT is often about a key concept...Good...Better...Best
It is not always up to us what we can build.