Virtualization Megathread V2: VMs inside VMs

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

«‹›2 »

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Martytoof posted:

Does VCSA use the same license as a host-based vCenter install? That is to say, if you have an Essentials license are you able to choose between one or the other, or is VCSA a separate license? Their Essentials Plus kit page seems a little vague about what the situation is.

I believe so. We have a normal "VCenter 5 Standard" license that works with VCSA.

cheese-cube posted:

Just killed our vCenter instance. Attempted to do a switchover in the Heartbeat Console and everything went to poo poo. Now the vpxd service is refusing to start.

Edit: I could really use a hand here if anyone has the time. The following event is being logged in the Windows Application event log:

Check that ADAM is running or that you can connect to your AD instance.

# ¿ Oct 19, 2013 19:08

Adbot: ADBOT LOVES YOU

# ¿ Apr 26, 2024 16:36

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

cheese-cube posted:

Thanks for the help PCjr sidecar! Any forums upgrades you'd like (i.e. Plat, Archives, etc.)?

No problem, glad I could help. Plat would be awesome!

# ¿ Oct 20, 2013 06:15

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

cheese-cube posted:

Cool, what's your e-mail address?

pcjrsidecar@gmail.com

# ¿ Oct 20, 2013 13:47

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Bhodi posted:

Dev has been creating hosts willy-nilly with no oversight and now vulnerability scans are showing findings with IPs and I need to discover if those IPs are actually virtualized servers in our enviorn or if I can pass the buck to someone else.

Check to see if the MAC addresses for those IPs are in the range assigned to VMWare (00:50:56); this won't work if the devs are setting the MAC address inside the OS.

# ¿ Dec 22, 2013 18:17

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

adorai posted:

Why does he disable the serial and parallel port in the bios instead of just deleting them from the VM? Seems like a waste for the hypervisor to emulate them just to disable them.

It wastes more time to do it that way, and more effort is obviously better.

# ¿ Jul 12, 2014 20:47

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

I'm fairly confident that the P410i is choking on the IO that you're throwing at it. Entry-level embedded RAID that would be fine with HDDs falls over badly at SSD IO rates. Consider using straight SAS controllers and add RAID in the VM (if possible) or add more controllers.

# ¿ Aug 14, 2014 22:21

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Martytoof posted:

That's exactly what I'm expecting, mainly because it's exactly what I got from HP about the hardware, but our vendor is putting HP and VMware on a bridge call so I'll let them duke it out

I 100% believe this is going to be as painful as I'm envisioning. I also entirely believe it's hardware based, but good luck trying to convince HP support of that.

What brand/model SSD? What percentage is in use?

in a well actually fucked around with this message at 03:05 on Aug 15, 2014

# ¿ Aug 15, 2014 02:53

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Martytoof posted:

I'm in datastore hell. The guy who set up all these hosts before me allocated 99% of the datastore to VMs, thick provisioned, with no thought for future expansion or backups. So now when I come in and am asked to do Veeam backups I am literally just .. "welp"

Also, with regards to my stupid SSD issue from a page or two ago. I booted up the loving VM today and my DAVGs are all normal. Worst -- loving -- timing since I just got VMware on the phone to start verifying this. HAS to be something to do with the SSDs, like garbage collection or something. Not even sure what to do now. I don't want to put my production VM back on that box in case it happens again but I can't find anything wrong with it at this point. My vendor actually verified that they were seeing the same problems I was so they'll back me up to a point, but I don't even really know what they can do about it at this point if everything checks out between HP and VMware.

How full are the SSDs?

# ¿ Aug 24, 2014 02:04

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

jaegerx posted:

There a normal post. Rrd files are just terrible in nature due to the small size and constant writing nature. Running them on lustre was a challenge. We tried gluster first but it wasn't up to it. We eventually found the sweet spot with 6 management nodes. We could've used fiber but the expense was too high so instead we use the fiber to support our vms over clvm so we can pass them around instantly via kvm.

This was with zenoss commercial edition. We needed some of the commercial packages to support our larger network hardware.

I've had good luck w/ rrdcached, if you can handle a few seconds delay for the data to coalesce. Lustre's really optimized around 1MB block writes so I can imagine RRDs on top it that being... interesting.

# ¿ Nov 29, 2014 19:13

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Does anybody have any good modern VDI best practices guides? Specifically, looking at hardware discussion (MHz vs core count, network config, etc.)

# ¿ Dec 9, 2014 21:55

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Wicaeed posted:

Does VMware offer something like KVM "Backing disks" (something my coworker calls them).

Basically they work as a stateless disk that allows multiple VM's to be spun up from a single master disk, but as soon as they are powered off, revert back to the original state.

Linked clones, kinda.

# ¿ Feb 11, 2015 21:29

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

BangersInMyKnickers posted:

Is there a good way to figure out how much overhead is being consumed by the iSCSI or NFS software initiator on a 5.5 host? I'm averaging around 4k iops on NFS with bursts upwards of 6-7k across 3 hosts in the cluster. I will be doing a hardware refresh in the next few months and am wondering if I am getting anywhere near the point where I should be messing around with hardware initiators or iSCSI offload with fancy emulex/qlogic cards instead of the more typical broadcom/intel stuff. I was expecting an NFS and iSCSI object under the System performance counters but nothing there jumps out at me.

In my experience, Emulex/QLogic Ethernet 'storage adapter / offload' cards are hilariously poo poo. If your 24-core 2.5 GHz server can't spare the miniscule CPU overhead involved, those cards are going to fall over at that rate anyway.

# ¿ Jul 3, 2015 03:10

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

cheese-cube posted:

Is this an OK place to talk about cloud virtualisation or is there another thread? Interested in hearing peoples experiences with Azure, specifically the new Resource Manager front-end. So far I've found some hilarious bugs with the Azure PowerShell 1.0.1 cmdlets (Mainly weird type handling and undocumented type restrictions for parameter arguments) and the Dashboard exposes some weird undocumented poo poo. Also if you're looking to deploy a Check Point VM with multiple NICs in RM then I may be able to help (Still not sure if it's 100% supported).

It's fine here, there's also http://forums.somethingawful.com/showthread.php?threadid=3702086 but it's pretty slow.

# ¿ Dec 14, 2015 17:23

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

The Enterprise / Enterprise Plus distinction has been bullshit since they rolled it out but lol at charging for the E+ upgrade.

# ¿ Feb 11, 2016 18:21

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Potato Salad posted:

What sort of claim does MS have on VMware IP?

When you have 30,000+ patents in your library it doesn't really matter.

# ¿ Feb 15, 2016 23:57

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Potato Salad posted:

Just a wide blunderbuss of expensive infringement claims?

There's a story about the dawn of the PC area.

Small startup gets a visit from a team of IBM lawyers. We noticed you infringed on our patent X. Interesting, the small startup says. Can we review these and get back to you? Followup meeting, a week later. Small startup says, We looked at the patent, and we don't use that feature. We've changed the way we do that so it's explicitly clear we're non-infringing. Senior lawyer says OK, pulls out a list of two dozen more patents that may infringe. Startup licenses, because IBM had enough patents and enough lawyers to find something that would stick eventually.

There are undoubtedly Microsoft patents VMWare is infringing on; thirty years of OS development gives you enough to draw on. Ultimately, everyone is infringing on everyone. The only defense is to have a big portfolio of your own.

# ¿ Feb 16, 2016 03:49

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

HPL posted:

near limitless computational power on tap

Why do you think that quantum computing would provide that?

# ¿ Feb 19, 2016 06:22

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Methanar posted:

Has anyone ever had ridiculous 350-500ms write latencies when using hybrid vsan?

Read speeds are fine, writing to SSD cache is fine, and the write latencies under vsan disk/deep dive show write latencies of around 3-10ms which is what you expect.

That makes me think the problem is between the vsan client tab and the vsan disk: ie the network fabric. But that doesn't make much sense either because clearly the data is getting to the servers just fine if the SSDs can write with zero latency. The hosts are connected together with 40g links with vcenter connected by 1g.

The m5210 raid card we're using isn't certified for vsan but is supposedly supported in raid 0 mode. Whatever it means. We interpreted it as meaning each drive had to be in a single-drive raid0 volume. Exposing single-drive jbod didn't allow vmware to recognize it, but the raid 0s were recognized, which sort of makes sense.

http://www.vmware.com/resources/compatibility/detail.php?deviceCategory=vsanio&productid=36877

Am I just doing something dumb and obviously wrong?

Entry-level RAID controllers generally don't play well with VSAN; that's probably the most likely suspect. I'd make sure you're following the steps in the linked KB in the compatibility page to the letter. Also make sure the firmware on your Mellanox cards are up to date.

in a well actually fucked around with this message at 03:02 on Sep 25, 2016

# ¿ Sep 25, 2016 02:59

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

If you have enough supercomputer to attack modern password hashes why aren't you running bitcoin on the supercomputer instead of scamming pennies from hijacked rackspace accounts?

(I originally said softlayer but what hackers are gonna set up a fax machine to provision new vms lol)

# ¿ Jul 23, 2017 16:33

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

evol262 posted:

Tooling took a while to get there with containers also. I'm not saying this is super practical yet, but the isolation offered by using virt is an active area of research. Tooling comes later, mostly.

They mimicked container workloads. Logging and telemetry in containers is handled from the host level. This is just more tooling missing.

To be honest, they should have compared to process isolation with bare cgroups, which docker (and all the others) is/are basically a wrapper around.

I'm basically taking this as "in 2 years, we expect that there will be a container-like solution which leverages virtualization for more isolation"

isn't that basically intel's clear containers?

# ¿ Nov 6, 2017 02:56

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

cheese-cube posted:

Is anyone here using VDP (vSphere Data Protection) and has found it to be an absolute dumpster fire? We've got it deployed at a couple of sites and it's a god drat nightmare. The appliances keep failing at random requiring reboots and/or lengthy excruciating troubleshooting. Also from poking around under the hood the whole "solution" is just complete trash.

On that subject what's the current go-to solution for snapshot-based VM backups that is free and isn't garbage? Our customer is demanding VM backups and is refusing to pay anything beyond the cost of implementing the solution. Normally I'd just sever but we're in a weird situation and need to provide something. So if I could replace VDP with something that doesn't randomly catch fire that would be swell.yewah

Vdp has always been pretty bad. We went to thread favorite Veeam and it�s good, so far.

# ¿ Feb 7, 2018 04:01

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

What is the guest OS set to in the VM settings?

# ¿ Apr 10, 2018 21:59

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

What�s the consensus on 6.7? Any gotchas coming from 5.5?

# ¿ Aug 26, 2018 17:41

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Number19 posted:

6.7u1 is coming later this year and has a feature complete HTML5 client so I would hold out for that unless you have a pressing need for another 6.7 feature

I have the gear for RoCE support on a new cluster, which was my primary interest in 6.7, but we can wait until U1.

# ¿ Sep 8, 2018 22:10

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Wicaeed posted:

Why does VMware make it so difficult to unmount a datastore & detach a LUN across multiple hosts?

Unmounting a datastore they've got down as bulk unmounting is fairly easy, but detaching the LUN is still a tedious task of going to each ESX host, finding the LUN and manually detaching it.

I've only found 1 or 2 scripts that can do the same thing, but they all jump right in and try to do things like unmount all datastores on a host, or they only do it for clusters of ESX hosts.

Guess I have to write my own, or something *sigh*

It is a feature in whatever version of esxi you�re not licensed for.

# ¿ Sep 23, 2018 02:12

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

TraderStav posted:

Just realized that there was a virtualization thread after posting this in the NAS/Homelab thread. Someone mentioned building a dedicated computer to physically attach to two sets of KVM that virtualizes but still connects directly, but my original inquiry was to do GPU pass throughs without connecting directly to the server for video/etc.

Would love some input on this, if I�m in the right thread, if not tell me to go away! Thanks!

This is the right thread as much as there is a right thread.

You�re trying to build a gaming VDI service on a five (?) year old server with a mishmash of old, slow GPUs.

Try to do it it might be fun if you like months of extremely fiddly challenges and want to learn a lot about linux virtualization and pcie and memory mapping and spend a lot of time with your children troubleshooting, but you are not going to get this stable by this fall.

# ¿ Apr 27, 2020 20:45

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

rufius posted:

Alternatively, the little NUC I bought (https://simplynuc.com/ruby/) sips power.

I'm not saying you should equivalence class them, but there are benefits especially in the realm of power:performance.

If electricity is cheap where you are, then never mind.

As someone who somehow got on their spam marketing lists despite never interacting with them or any related area and have been unable to get off of it, gently caress them.

# ¿ Aug 19, 2021 17:48

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

A container is two things:
1. A turbocharged chroot, but for all the resources (network, process tree) instead of just a filesystem, using namespaces and cgroups. Uses the same Linux kernel as the host (usually).
2. An image based software distribution standard.

LXC, Podman, Docker, etc. all do the same kind of things for #1. There�s a lot of (oss) drama around them, but it doesn�t really matter for the same basic processes (download image, start/stop process, manipulate the namespaces.) If you�re happy with podman there�s no need to look at lxc.

Because containers normally run isolated you need to provide all the executables and libraries to run within. This is done via binary images. This can be bone stock debian/rhel, or container optimized os spins that provide just the binary statically linked with the relevant libraries, or some point in between. Alpine Linux is frequently used as a lightweight option.

For standard stuff you get your container from dockerhub or other container repository. It�s (usually) a series of gzipped overlayfs images that docker pulls. It doesn�t really matter as your tool will manage it.

Images are built from Dockerfiles. Basically a rpm spec file.

Containers are usually* immutable.

As jackal mentioned, kvm is standard virt, not emulated. qemu provides the fake hardware (pci etc) but doesn�t do cpu emulation. The rhel tools are fine.

# ¿ Apr 12, 2022 17:19

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Yeah. When the vultures arrive it�s a good time to get gone, for both employees and customers.

# ¿ May 28, 2022 01:22

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

The world will always need VMware admins in the same way the world will always need mainframe admins. Whether that�s how you want to spend your career is up to you.

# ¿ Jun 30, 2022 13:50

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

skipdogg posted:

I see no lies in what this guy wrote. Apologies for the linkedin article but that�s where he posted it.

https://www.linkedin.com/pulse/brian-maddens-brutal-unfiltered-thoughts-broadcom-vmware-brian-madden

Reposting this for a good read on vmware�s current and future market position.

# ¿ Jun 30, 2022 13:51

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Mr. Crow posted:

That's a dumb take and there are still and will always be a lot of use cases for staying on-prem

There may be lots of use cases for staying on prem, but fewer that require VMware. Grudgingly maintained reliance on enterprise tech from a vendor focused on extraction-maximising subscription revenue while planning on reducing headcount in development and support is a common pattern, and not one I�d be excited about building my career around. It�s not going anywhere; I�m sure people make good money running Oracle, too.

# ¿ Jun 30, 2022 16:24

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

I mean, you could read that blog post; you could read Broadcom�s investor PR about increasing revenue and reducing costs, or you could look at the trajectory of enterprise platforms (enterprise- not dominant tech companies; NVIDIA isn�t selling P100s and charging a subscription for cuda.)

Their core products are stagnant in tech and in revenue. They may be dominant now, but IBM was dominant in the 80s and Oracle in the 00s.
No CIO wants to pay Oracle or IBM or SAP, but they do, and they�ll pay Broadcom.

It�s not going away, but I wouldn�t advise a generalist to investing in developing skills in it, because I personally think things that look like cost centers or have ongoing maintenance that aren�t tied to revenue are less valued by leadership and less compensated, particularly when there�s a fountain of cash for highly transferrable skills in other areas.

# ¿ Jun 30, 2022 18:30

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

Subjunctive posted:

I�ve got a small Ubuntu home server running Home Assistant and a Unifi Controller under a creaky qemu/kvm setup, plus some other under-maintained services. I also have a Steam Deck which is shockingly compatible with the things I�ve been playing. I have virtualization questions about both arenas!

For the home server, I think I�d like to get rid of the mini tower and move to a NUCish form factor, and beef it up so I can play with some more modern homelab/clustering things. My light reading has led me to a Simply NUC Ruby r8 on which I will stick proxmox and then figure out how I want containers and VMs to interplay. And then stub my face on k3s, probably. Is this a sane path to pursue?

For gaming, the time for my Zen 4/Lovelace upgrade is coming and I�m seriously thinking of giving Linux gaming a shot for the first time since �Civ: Call to Power�. If it doesn�t work out for everything, I will probably want to do GPU+USB passthrough to Windows 11 or similar. Can I plausibly do that with a single NVIDIA GPU? I�ve heard tell, but I don�t know how reliable it is.

What�s the state of the art for doing things like clipboard in/out of VMs? I�ve only used the VMware stuff for that, and I�d rather not entangle myself with their stack if I can avoid it.

Thanks for any guidance you can provide!

Don�t buy from SimplyNUC; I got on their mailing list a decade ago and they�ve been sending me spam constantly since, and none or their unsubscribe links work. They also change domains and senders around so it�s annoying to block.

gently caress those guys.

# ¿ Oct 1, 2022 14:53

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

SlowBloke posted:

In case you guys missed the memo, Broadcom woke up and chose violence.

https://news.vmware.com/company/vmware-by-broadcom-business-transformation

You won�t be able to renew SnS on perpetual contract beyond the current duration and won�t be able to purchase new perpetual licenses. Conversion to subscription might also be the time VMware moves to cores instead of sockets but who knows.

Yeah they�re doing exactly what they said they�d do.

# ¿ Dec 14, 2023 14:14

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

BlankSystemDaemon posted:

It's not just what they said they'd do, it's what everyone else said they'd do before they said they'd do it.

New speedrun category: Become Oracle at any% cost!

IIRC when they made the offer public they were very explicit about their plans for cutting costs (staff) and increasing revenues (prices).

Oracle at least still puts effort into development on their various products like ExaData or their cloud platform; Broadcom is in classic enterprise resource extraction mode.

# ¿ Dec 14, 2023 20:43

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

fresh_cheese posted:

I have briefed multiple new customers planning to buy new mainframes specifically because it will be cheaper for them to port their linux junk to s390x and move to kvm than to stay in vmware with the expected pricing change

I wonder if that increases or decreases now that the change happened

2024, the year of Linux on the mainframe.

# ¿ Dec 15, 2023 15:54

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

fresh_cheese posted:

dont want it to do anything cool or good.

You already said openstack, no need to repeat yourself.

# ¿ Dec 17, 2023 16:34

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

CommieGIR posted:

Puppet and Chef support multi-hypervisor setups and you can script it.

Did you just tell me to go gently caress myself?

# ¿ Dec 17, 2023 16:42

Adbot: ADBOT LOVES YOU

# ¿ Apr 26, 2024 16:36

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

HalloKitty posted:

Is this the gently caress Broadcom thread?

I'm not at all looking forward to the Hyper-V and SCVMM that is without doubt in my future.

Broadcom wants $FUCKYOU more each month for VMware. We have Datacenter licences for Windows already, but we'd rather pay for VMware on top and use that as our hypervisor than Hyper-V, but that's inevitably coming to an end with these stupid price increases.

I don't really get the play here - if all SMBs switch to alternatives, it will not only hurt stable development of VMware (few customers = fewer testers hitting corner cases resulting in a less battle tested product), but also future adoption in general - if everyone starts out with Hyper-V, Proxmox or XCP-NG or maybe something else, then in future they'll go with what they're comfortable with. It's why Windows is so widespread; Microsoft got it in the classrooms around the world as best as it could. VMware is also learnt by many today, but that could change overnight, and probably will. Within a short time VMware will be a dead product and company.

vv Yeah, it's honestly not a bad idea.

The play is that it is not a growth opportunity anymore. They�ve captured all the market they�re going to and it is not easy to switch. It is now legacy enterprise software and the pattern around that is well established.

# ¿ Feb 12, 2024 17:09

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

«‹›2 »