Virtualization Megathread V2: VMs inside VMs

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Virtualization Megathread V2: VMs inside VMs

«‹›9 »

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Bob Morales posted:

What about running something like a firewall appliance on a virtual machine, as long as it has a dedicated set of NICs?

We do it without dedicated nics, just use vlans.

# ¿ Sep 18, 2012 13:10

Adbot: ADBOT LOVES YOU

# ¿ May 9, 2024 14:40

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

evil_bunnY posted:

Apple doing fuckall regression testing? Incredible!!!

Why would they need to test third party apps? All you need are apple products. I'm just surprised they didn't just announce and release something named iVirtualBox today.

# ¿ Sep 22, 2012 01:42

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

I don't personally see a problem with that, but unless you are using a nexus 1000v, each physical nic that they intend to use storage for should have a seperate vmk for iscsi specifically bound to it.

# ¿ Oct 18, 2012 01:46

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Erwin posted:

How much time skew do you guys see between VMs? I didn't really think much of it, but our programmers have reported seeing up to 10 seconds between different VMs, causing small issues like builds not triggering on continuous integration servers if you quickly check in code in rapid succession. This is a mostly Windows environment, and everything syncs to a DC which syncs itself with an external NTP server. The DC is virtual.

We use group policy to specify windows server time sync should use an NTP server rather than DC. Specifically, a vbscript runs on startup to set specific registry keys. Below is an example of the code for the NTP settings.

code:

Dim Path
Path = "HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\W32Time\TimeProviders\NtpClient\"
set ws = WScript.CreateObject("WScript.Shell")
o=ws.RegWrite(Path & "SpecialPollInterval", "900", "REG_DWORD")
v=ws.RegRead(Path & "SpecialPollInterval")
Set WSHShell = nothing

Path = "HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\W32Time\Parameters\"
set ws = WScript.CreateObject("WScript.Shell")
o=ws.RegWrite(Path & "NtpServer", "time.domain.local", "REG_SZ")
v=ws.RegRead(Path & "NtpServer")
Set WSHShell = nothing

Nothing special, just syncs the time to our time server (time.domain.local) every 15 minutes. Since doing this our time has been much more consistent than it was back in the day when it synced to our DCs. All of our DCs are virtual, and they get their time the same way. Our NTP server is also virtual, but polls it's NTP servers every minute or so.

# ¿ Oct 22, 2012 21:19

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Rhymenoserous posted:

This wouldn't have happened if you ~~owned a gun~~ had vmware

yeah when weird poo poo happens in VMware it just kills the host management, so you literally cannot do anything. You are then forced to choose between killing the host with all the VMs on it and letting HA do it's thing, or shutting down every VM on the host one by one.

I'm not saying Hyper-V doesn't suck, just that VMware has some show stoppers as well.

# ¿ Oct 22, 2012 21:28

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

mstsc for life.

# ¿ Nov 1, 2012 22:54

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

I have two netapp HA pairs at seperate sites. I have two vCenters each with a datacenter, and the vCenters are linked. I have SRM installed. I have created protection groups, have my SRA configured,etc.. When I test my SRM failover, my datastores (which are NFS) are cloned properly but are only mounted on a single host at my DR site. This is less than optimal. Does anyone have any idea where I should start looking to determine why my other hosts are not mounting the datastores as well?

# ¿ Nov 3, 2012 23:09

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

E4C85D38 posted:

With my test environment at home I'm hoping to pick up some virtualization knowledge, so I figured virtualizing the existing domain controller would be a good exercise. I can't seem to find this in documentation, but: if I use vCenter Converter Standalone to throw an existing physical machine on a USB external drive, install the free VMware vSphere Hypervisor on the physical machine�, and then plug the USB drive into the server, will I be able to import the VM on that drive to the hypervisor or do I have to do something silly like SCPing it over the network?

_{� First making an image of the existing OS if things go wrong, of course.}

ESXi will not recognize usb storage.

# ¿ Nov 4, 2012 00:59

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Corvettefisher posted:

Yeah can't say I have seen that issue, porbably because I use FCoE/iSCSI for most things. NFS permissions might explain it.

It wasn't permissions, it was even dumber. By default, the maximum number of NFS volumes that an ESXi host can mount is 8. This was fine for our daily volumes, but when you double them, we needed more.

It didn't affect the one host because we had temporarily used it in our production environment for a week, and had already bumped up the max nfs volumes to 32 on that one host.

# ¿ Nov 4, 2012 19:33

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Rhymenoserous posted:

That or go to a worker server paradigm, but at that point you may as well go bare metal and yes I'm flogging myself for saying that.

It's also not true. In this specific type of example, I might actually prefer to buy an 8 core box and create a 7 core VM that will run 100% than to go bare metal. Reason being that abstracting the hardware away from the application makes it easy to quickly upgrade the hardware with effectively no change to the guest except more speed.

# ¿ Nov 21, 2012 07:04

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

KennyG posted:

I would love to utilize this on the host as I coud sub-divide a 800GB drive into 400GB for prod and 50-100gb for various environments and greatly improve flexibility. However can I do this with shared storage and v-motion?

setup an openindiana or freebsd guest on the host with the fusionIO and share it out via iscsi or NFS. Add that disk to your guest OS. Beware: if that host goes down, the datastore will go with it, so I would recommend exporting from two seperate hosts and using software mirroring or something. At that point, it may not actually boost speed.

# ¿ Nov 21, 2012 07:07

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

KennyG posted:

How would this work. If I did this why would I share each through the hypervisor/network instead of deploying vms with scratch disks directly?

You would lose your ability to vmotion and leverage VMware HA.

# ¿ Nov 22, 2012 04:48

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

just use vmware converter, it will solve all that ails ya.

# ¿ Nov 28, 2012 01:15

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

evil_bunnY posted:

I've seen people defrag VMs that lived on copy-on-write SANs. For them too, there is a special place in hell.

I've never seen it, but I've heard of people who deploy VDI solutions and don't turn off any automatic defragging. As I understand it, that is bad.

# ¿ Nov 29, 2012 01:12

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

KennyG posted:

How hard is it to re-mount a lun/vm and move it from one cluster to another.

Just add the lun to the second cluster. No need to remove it from the first until after you move your app.

# ¿ Dec 3, 2012 04:56

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

madsushi posted:

Don't quiesce the data. You're just going to have awful performance during the VMware snapshot creation/deletion and it doesn't buy you anything at all.

Let NetApp take the snapshots at will (via VSC or a snapvault schedule) and then replicate it like that. You get a 'crash-consistent' backup that is going to work. Can you remember the last time that a VM failed to come up after you did a hard power/reset on it? The answer is 'never'.

You are a lot more brave than I am sir.

We snapshot and replicate our NFS volumes 6 times per day, plus one additional that quiesces the VMDKs. It keeps my (paper) RPO low while giving me the warm and fuzzy feeling of a real backup once per day. We can afford the IO hit while 99% of the company is sleeping.

# ¿ Dec 6, 2012 05:10

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Misogynist posted:

Are there any particular gotchas with using VMware Converter on a Windows 2000 host these days?

pretty much whether the random number generator is rolling in your favor.

# ¿ Dec 7, 2012 06:03

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

based on what I've read, we are waiting for the next release to upgrade from 5.0.

# ¿ Dec 14, 2012 18:01

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

goobernoodles posted:

Does anyone have any experience with Vmware Site Recovery Manager?

My environment currently consists of two offices:

Main Seattle office: 4 ESXi hosts, virtual center server, 12 VMs, iSCSI SAN
Portland Office connected via MPLS: 1 ESXi host running 4 VMs on local storage.

I'm hoping to implement some sort of two way SAN replication so that we have SAN's in both offices backing up each other. Currently we're paying a 3rd party company to move our BDR backups off-site and it's not exactly cheap or backing up at the VM layer.

We run SRM quite successfully. It works as advertised, and is pretty easy to configure.

# ¿ Dec 20, 2012 00:05

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

cheese-cube posted:

Are those one of them Cisco UCS servers (I can barely make out what looks like a Cisco logo on the left bracket of each server)? They look strikingly similar to IBM System x rack-mount servers, including the hexagonal vent holes which I'm pretty sure IBM has some stupid patent on.

Oh and yeah a server thread would be awesome, I'd gladly contribute a ton of info on the IBM side of things.

Yes those are rack mount Cisco UC's servers. If you have a relationship with your Cisco account rep you can just threaten to buy HP instead and they have a lot of room on the price.

# ¿ Dec 20, 2012 13:27

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Corvettefisher posted:

That's a UCS for you, each of these were probably in the range of 24k each.

I am not sure how you optioned those, but we were looking at around $10k each for dual e5-2670 procs, single 10gbe nic + onboard 1gbe, and 96GB of ram. No disks. This was in april.

# ¿ Dec 20, 2012 19:28

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Corvettefisher posted:

Not to different however mine have a DAS setup due to the environment the will be going to. That was MSRP, I am not sure what sales does with it and brings it down to.

I would assume it is 44% off MSRP for hardware and 15% for smartnet. Maybe 48% if you were buying enough of them at once.

I only say that because I don't want people in this thread to hate on UCS for the price when it really is a very nice platform, and in reality the cisco tax isn't that high on it. The seemingly high price includes a >$1k 10Gbe NIC, a lot of RAM, and their ILO product is included, not an extra $500. A nice surprise when we bought ours is that they include sdcard slots on the motherboard, and ship with an 8GB sdcard in them for loading ESXi onto. It's probably $20 total in additional cost but is very nice to have.

# ¿ Dec 20, 2012 22:44

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

KS posted:

That's a typical discount for Cisco network gear, but UCS discounts are actually quite a bit higher. We're not really supposed to share specific percentages like that, are we? The whole IT purchasing thing (deal registrations especially) seems incredibly dishonest sometimes.

I certainly never signed any kind of NDA. If a vendor wants to play the stupid discount game instead of just giving me a price that he makes a small profit and I don't get bent over, I have no problem discussing the discount.

three posted:

Have fun getting ripped off on Smartnet even if you get UCS cheaply.

Three years was somewhere around $1000/chassis if I recall correctly.

# ¿ Dec 21, 2012 01:00

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Fancy_Lad posted:

To be honest, XenServer has enough stupid quirks that I'm cautiously optimistic about the change. Mostly likely I'll just get to start bitching about stupid Microsoft quirks instead tho

Please tell me more. I am considering a 6 to 10 host xenserver deployment for a 700 user VDI initiative, and really don't want the spend the money for ESXi processor licenses when xenserver is "free". We are already licensed for XenDesktop, so going with View is not an option.

# ¿ Dec 22, 2012 00:21

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Fwiw we have the same problem with some of our xenapp guests on esxi 5.

# ¿ Dec 22, 2012 14:28

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

We are looking at replacing our ample Xenapp deployment with a xendesktop deployment running on xenserver. Does anyone have any experience with running xendesktop deployments on amd bulldozer chips? The 16 integer cores per proc seem very appealing for the application. As a follow up question, is there any kind of api for xenserver like vshield for vmware?

# ¿ Dec 28, 2012 00:57

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Corvettefisher posted:

I have a bulldozer right now actually, I am putting together another whitebox ESXi server. However testing view, not xen app; but I'll let you know how it goes!

Here is a nice review of the bulldozers by anandtech.
http://www.anandtech.com/show/5058/amds-opteron-interlagos-6200

We almost went with bulldozer for our main VMware farm, but we had a single application that didn't play well. it was important enough to sway us away. For VDI though, it's a whole different animal so I really am looking for real world experience.

# ¿ Dec 28, 2012 01:05

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Corvettefisher posted:

To the people asking about SRM
http://www.vhersey.com/
My teacher has a really great writeup on SRM if you want to check it out. Home Lab and all

The reality is that SRM is so easy to setup and configure, a lab is more or less the entirety of the training you need. If you have any experience with VMware at all, you should be able to install and configure it in a few hours with no experience with SRM itself. In fact, I would say licensing it is more complicated than implementing it.

# ¿ Jan 3, 2013 00:47

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Corvettefisher posted:

I wouldn't really consider a dual controller single chassis a HA deploy. A fire of a PSU or controller would trash the box, then I again I could be over analyzing it a bit.

Really?

A fire is a disaster. a fire in your datacenter is going to kill a fuckload more than your SAN. The fire suppression system is going to kill all of the power, flood the room with some kind of fire suppression gas, and evacuate the entire building until the fire department clears reentry. You had better be prepared to activate your DR plan for any kind of fire in your datacenter.

# ¿ Jan 8, 2013 23:51

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

OnceIWasAnOstrich posted:

Sorry if this is a dumb question. I'm trying to use ESXI 5.1 to virtualize a couple of machines we use in our research lab. I've got two physical NICs. I originally put the management interface on vmnic2, using the IP 10.1.185.4, and everything is great. I wanted to add a VMKernel port so I can hook this up to our NAS with ISCSI. I add a VMKernel port to vmnic1 and connect using the IP 10.0.185.2. Suddenly my client loses connection, and I can no longer connect to the management interface at 10.1.185.4...but I can connect to 10.0.185.2.

Have I done something horribly wrong, or is it intended that this happen?

remove the default gateway on the interface for your storage.

# ¿ Jan 19, 2013 02:06

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Corvettefisher posted:

VCAP-DCA Scheduled for PEX let's see what this does

It'll get you a job offer that you turn down.

# ¿ Jan 26, 2013 05:06

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Powdered Toast Man posted:

So I assume that the latest ESXi is pretty good at co-scheduling...that being said, is it a bad idea to have more than twice as many vCPUs on guests as you have actual logical processors on the host?

We are somewhere around 5 to 1 on our server farm.

# ¿ Feb 2, 2013 05:01

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Internet Explorer posted:

But if you are overcommited 5 to 1 because each one of your virtual servers has the same number of vCPUs as your processors have cores... that may be a problem.

We are very conservative and only provision one vCPU per box unless we experience a need for additional. We regularly ignore vendor requirements for more than one. We have 6 8 core processors in our main server cluster and approximately 200 VMs. Of those 200, roughly 1/4 have 2 vCPUs, and a few have 3 vCPUs. We have 128GB of RAM per processor.

# ¿ Feb 2, 2013 06:04

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Erwin posted:

I've already configured emails, that's not the point. You're not helping my impotent ranting.

I recommend a decommissioned server and a hammer. Sledge or ball peen, either will be fulfulling.

# ¿ Feb 6, 2013 00:10

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

We thin provision everything. We also have multiple sets of eyes on our alerts, which are properly configured.

Thin VMDK on top of thin luns on top of thin volumes.

# ¿ Feb 7, 2013 05:28

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

GrandMaster posted:

I started building our new cluster yesterday - 16x Dell M620 2xE5-2670 / 256GB RAM
Customized ISO & kickstart scripts saved my life, the cluster was built in a day! Would rather gouge my eyes out than configure the same 20 portgroups on each host.

use a dvswitch, takes like 5 minutes for us to configure networking on a host.

# ¿ Feb 9, 2013 00:03

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

evil_bunnY posted:

Yeah really, you just need (maybe) one vSwitch for management.

Strongly recommended. I was a dumbass once and changed the VLAN on our production vSwitch (thought it was DR side) for the management network. It was fun getting back into it.

# ¿ Feb 9, 2013 00:15

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Aniki posted:

I have old phone server software that I installed on a Windows 2000 Server VM using Hyper-V on Windows Server 2008 R2. The software has a hardware key that plugs into a parallel port. We forwarded the port to the VM, however the VM is not recognizing the hardware key. I noticed that one workaround that people are using for USB hardware keys is to plug them into an Network-Attached USB hub. Do you think that I could get away with plugging a USB to parallel adapter into a Network-Attached USB hub or is there an easier way to get the parallel port hardware key working? I do have a parallel card in the server. Also, would running Hyper-V as an administrator make any difference? Does VMWare have better support for parallel ports?

I'm not too experienced with virtual machines, so sorry if some my questions are a bit basic.

Thank you in advance for any help.

In my experience getting hardware licensing HASPs to work with a USB to parrallel adapter is hit or miss, but if it works with the cable, it should be trivial to use the USB to LAN hub. We do it and other than it occassionally needing to be reconnected, it's been pretty good.

# ¿ Feb 12, 2013 01:19

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Does anyone here run KVM in a serious business environment?

# ¿ Feb 14, 2013 04:36

Adbot: ADBOT LOVES YOU

# ¿ May 9, 2024 14:40

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Less Fat Luke posted:

Did you have any specific questions, or were you just wondering if anyone does it?

really just looking for the warm and fuzzy that someone actually does it.

# ¿ Feb 14, 2013 13:37

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Virtualization Megathread V2: VMs inside VMs

«‹›9 »