Hi all,

I'm hoping to get some (wise!) opinions on what I should or could do here.

I've got a 7 year old HP SAN that's nearing it's capacity. It's connected to a HP DL360 G7 running an old version of openindiana which operates as our fileserver. The SAN doesn't host VMs or databases, just home and shared directories for file storage. Now rather than try to increase the capacity of this SAN, I think it would make more sense to move our file's onto a new platform.

I know virtualisation is a must and we run VMware here so most likely I'd do something like this

1. get a newer HP DL360 or Dell RX730 or something similar
2. install ESXi 6.5 onto SD card
3. Setup a Raid 10 array using a hardware raid card.
4. Install a VM with one vmdk for the OS and one for the storage area.

Now I'm looking for advice on what the best OS and filesystem to use for the file server VM? Capacity would need to be in the region of 8 TB. We're almost entirely a linux operation here though we have a few windows clients. The current file server uses NFS v3 and samba for the fileshares, and the filesystem is ZFS. I'd like and have been asked, to keep the snapshot functionality that ZFS offers.

So one option is to stick with Openindiana and ZFS. One problem here is that it's an OS I'm not so familiar with. However it's been pretty solid for us except that once the storage pool where our home directories live gets close to full then our workstations start to crawl.
I could go with BTRFS as a filesystem to keep the snapshot ability. Am I best to use one of Opensuse or Suse enterprise server in this case?
Is there an issue using either of these filesystems on top of hardware raid? I guess not if I'm not using their RAID functions?
Centos and XFS with LVM? Do LVM snapshots provide similar functionally to the ZFS ones?
Or stick with ZFS but use FreeBSD.
Or finally just get a Synology and use that.

I'd be grateful for opinions including if it's just to say definitely don't do one of the options for some obvious reason that I'm not seeing.

cheers!

Thanks!
Long time listener - first time caller!

I'm sysadmin for a small software company. It's a great job but covers so many areas that sometimes I feel I'm stretched a little thin keeping up with everything. I do tend to visit here quite a lot. I almost always find interesting and new IT topics to read...

I've got the Huawei P20 Pro.
I've had it for 9 months or so. Previous phones were a Samsung Galaxy S7 Edge which my 2 year old son threw down a stone stairs and a OnePlus 2 which he threw into the toilet. I can honestly say the Huawei phone is the best I've had by distance. The camera setup was a step above the other falgships at the time and I believe the P30 continues that trend. When my son finds a way to destroy this one I'll definitely be looking at Huawei again, if they're not completely crippled by the Google stuff.

In the interests of "closure", the plethora of libvirt updates which came down the line yesterday from fedora seem to have fixed the problem. I can now create virtual networks again. Yay!

Congrats Scott. Is there any extra benefit to you if I buy from Packt Publishing rather than Amazon, or vice versa?

Congrats! and Thanks!

We're running ESXi essentials here. One reason I'd favour sticking with VMWare is that it's what I know best. Another is I'm kind of a 1 man band here as a sysadmin but I do have a deputy who works a tester who I've trained up a little bit in ESXi in case I'm not around and something happens.
Yes we're on ESXi 6.5 so it'd be a VMFS 5 datastore.

@scottalanmiller Sorry, my choice of terminology didn't make things very clear there. By replicating I mean replicating the functionality rather than the snapshots themselves. At the moment if I developer deletes a file by mistake they can go to the .zfs folder in the root of their home and restore the file from one of the daily snapshots. I think if they were no longer able to do the equivalent of this, that I might get some resistance.

@scottalanmiller said in OS/Filesystem for file server ~ 8Tb:

@Doyler3000 said in OS/Filesystem for file server ~ 8Tb:

We're almost entirely a linux operation here though we have a few windows clients.

Just have to mention that this fact makes ESXi a weird choice as you need a range of skills instead of being able to focus on the ones that you have. If you have Linux skills, you are ready for KVM or Xen

In my (shaky) defence I had stronger VMWare skills than Linux skills at the time

Thanks all, I very much appreciate lots of great advice.
I've already got a XenServer installed and I'm going to take a look at Xen Orchestra now. As regards OS/Filesystems OpenSuse (or Centos) with XFS and LVM will get some testing in the near future.

Out of roughly 30 VMs, 2 are Windows server instances. We're mostly running Centos here.
I'll look into SA licensing. I'm not very familiar with it.

@scottalanmiller said in Hyper-V replication, Starwind, or something else?:

If Hyper-V is the right choice here, that's not a surprise. And if it is because it is part of the appliance, in a lot of ways you can ignore it. A server with a hypervisor is not exactly the same as an appliance with one. The appliance is a "black box" in theory and how it does its magic is of no concern, until it is exposed to you, if that makes sense.

Yes that's how I've been trying to think about the appliance. Also since it comes with Active Support from Starwind, in theory my worry about less technical people having to support this when I'm away becomes less relevant. In fact, theoretically I would barely need to support it

@sanwin

Actually am I talking rubbish there? The hypervisor will still be Hyper-V so the VM's theselves won't require any conversion?
Just the VSAN software itself will be running in a Linux VM rather than a Windows one.

We've recently started to use a Meeting Owl in our smaller meeting room (maybe 6 people capacity)
The remote workers love it and insist we use it rather than a traditional front of the room cam.

https://www.owllabs.com/

@JaredBusch We're using Zoom.
We had been using something called Starleaf but it didn't work with the Meeting Owl as well. That's one of the reasons we moved to Zoom.
There's a Meeting Owl Pro for bigger rooms but it's not available for us in the UK until later this year. When it is I think we'll be buying that for our main conference room.

@scottalanmiller said in Reconsidering ProxMox:

@stacksofplates said in Reconsidering ProxMox:

@stacksofplates said in Reconsidering ProxMox:

@scottalanmiller said in Reconsidering ProxMox:

@stacksofplates said in Reconsidering ProxMox:

After all of this, I still don't get the use case for LVM backed VMs. Other than possibly, possibly a super IO heavy database. Even then, it's questionable.

That's roughly it, and yes, it remains questionable at the best of times.

In the cases where you need LVM fat, you almost certainly also need to avoid LVM because that itty bitty overhead is still too much.

Preallocated qcow2 images are 99% as fast as LVM volumes. Even with just preallocating just the metadata I've had almost native disk write speeds. You lose all of the advantages of qcow2 like libguestfs, the qemu agent, internal and external snapshots, etc.

that said, no idea how the eff you do that with ProxMox. That was just KVM.

It's the default actually. We use Qcow2 on LVM-Thin mostly.

Hi Scott (and everyone else),
I've been playing around with Proxmox for a week or so. I haven't used LVM thinpools before so I wanted to check if I'm making sense here. Proxmox doesn't let me put a qcow directly onto a thinpool (like the local-lvm created by default).
Do I need to create a volume group on top of the thinpool, and mount that as directory storage to be able to use qcow2 on LVM-Thin as you're doing?

Cheers!

Well since I use this forum a lot and I don't want to be seen as 'weird' I'll try to explain where I was coming from.
Yes I've been using vanilla KVM for a few years and I've always used qcow2. There was some talk about qcow2 earlier in the conversation with stacksofplates being quite positive about their advantages and then SAM saying that he uses qcow2 on lvm-thin mostly and it was the default for Proxmox.
So I thought - those guys generally know what they are talking about, let me see if I can make a qcow2 on lvm-thin.
But proxmox wouldn't let me until I created some directory storage on the thinpool.
Ok great that works - but that seems a little strange - let me ask the guys and gals on the forum what they think.
And here we are.

As regards moving to an appliance - I'm comfortable with KVM but I run our VM infrastructure mostly on my own. On occasions where I'm on holiday or otherwise unavailable, it would be helpful if at least a couple of others could administer the system (particularly backups).