Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
KennyG
Oct 22, 2002
Here to blow my own horn.
I have read the past 35 or so pages trying to get up to speed, but I have one specific question. Does VMWare support clustering at the guest level.

e.g. If I want a 8 core guest, does that absolutely require a host with 8+ cores? More importantly can I cluster cheaper quad core single socket commodity boxes instead of springing for the pricier 2 or 4 socket boxes?

I have some apps that do well on a threaded basis but don't really split up into multiple VM's very well. Think like video encoding. I would like the ability to quickly spin up a VM with a lot of available resources and then shut it down when the batch is over (overnight window).

Adbot
ADBOT LOVES YOU

KennyG
Oct 22, 2002
Here to blow my own horn.
To be clear - we have some multi-threaded applications that do a batch process. I can't say what it is as it's proprietary but let's take the annalogy of video encoding because it's common and people understand it.

Basically, you can encode a file F in X/T time where X is the encode time and T is the theads. However, if you have A) pair of single socket quad core machines and b) one machine with 8 cores, A) can produce two halves of an output in the same time as B) can produce 1 complete output set. Since re-combining the output into a complete set is a non-trivial task, having a machine that is as large as possible results in the smallest time for the desired compute task.

The rub is that pricing of vCenter is effectively linear, by processor. If I install it on one 4CPU box it's the same as installing it on four single cpu boxes. However, Hosts are close to geometric by processor. The 4P tax is pretty nasty. You also can't even buy an 8 socket host. All of this means that it can be more cost effective to buy 8 x 1 socket systems than buy 2 x 4 socket systems.



Thank you KS. That answers that - so it's better to eitehr get bigger hosts or re-code the application to be more modular. I can't really say which it will be but I'm guessing 4slot hosts are cheaper than re-development Thanks.

KennyG
Oct 22, 2002
Here to blow my own horn.
I doubt 32 will remain the limit, but at that time maybe VMWare will implement the feature too. Hardware is an expensive sunk cost, where on software you usually can get a sizable discount on full retail with the upgrade price. (I can't specifically speak from experience from VMWare but I would be shocked if it weren't true.)

It's a tough sell to the senior leadership to ask for $$$,$$$.$$ in hardware and then turn around in 2 years and ask for it again. Our cycles tend to be more in the 5-7 range, even for servers.

I have tried pointing out that we are paying more in power than to buy new hardware but they just think I want new toys.

As to the AMD question, I have seen several people talk up the AMD's in CPU heavy virtualization and perhaps the 6300 series will be better but at this time it seems like a rather sizable political gamble on my part.

KennyG fucked around with this message at 23:47 on Nov 20, 2012

KennyG
Oct 22, 2002
Here to blow my own horn.
Ok, another question I forgot to ask that came up when I was reading.

We have an Oracle App ( :smuggo::respek::smuggo: ) One of the many features of Oracle is a flash cache space. This is essentially a scratch disk that is used like an "L2 cache" of sorts for things that would have aged out of the buffer cache and back to the hard disk. By design this is put on some extremely low latency storage space like a SSD or better yet something like a fusionIO IO drive. .01 ms IO latency vs 4-10ms That's a pretty good boost.

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?

We have a NetApp iSCSI SAN and I obviously want to keep the benefits of shared storage, but since Oracle effectively uses the storage space as volatile storage (it does not count on that data being there when it reboots) I was wondering how I could expose it to the guest and make it work.

Assume identical hosts in a 3 machine cluster. What buzzword/feature am I looking for? Oracle simply maps the space to a path in the guest OS filesystem.

KennyG
Oct 22, 2002
Here to blow my own horn.

adorai posted:

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.
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? Wouldn't putting the traffic back to the network eat the benefits of low latency io.

Unfortunately 800gb of ram is about 4-5x the price of a enterprise IO drive. Given our storage /vm requirements I'd likely go local storage if I can't do hybrid shared storage unless I am misunderstanding what you are saying?

KennyG
Oct 22, 2002
Here to blow my own horn.

adorai posted:

You would lose your ability to vmotion and leverage VMware HA.
Ok that's the why but what about the how?

Doesn't sharing it out with nfs or iscsi through a guest across the LAN or vLAN take latency a from a few micro seconds to milliseconds?

KennyG
Oct 22, 2002
Here to blow my own horn.
How hard is it to re-mount a lun/vm and move it from one cluster to another.

Basically, I'd love to be able to have a 3 machine cluster for this Infrastructure app we have but the oracle license per core is just loving killing me. If I separate the cluster and remove the ability to run it on other machines, I would only have to pay for 1 box and would be covered for up to 10 days a year of disaster.

16 cores at ~$25k a core is $400k - if you put that on 3 boxes, now you're looking at effectively paying $800k for vMotion/HA. gently caress larry Ellison.

KennyG
Oct 22, 2002
Here to blow my own horn.
Should I be concerned if my integrator/rep still thinks that I need vSphere Ent+ since I wanted to have 192GB of ram per host?

I freely admit that I know next to nothing about vSphere or any virtualization issues, but I don't want to blindly follow some idiot who doesn't know what he's doing.

It's Dell, btw. (Fed Gov, before someone starts trying to poach buisness.)



Second question, can you virtualize your vCenter Server within your cluster? Is that a good idea?

KennyG
Oct 22, 2002
Here to blow my own horn.
I guss virta/HA/affinity for vCenter is better than buying a whole host and dedicating it. 4 host/~40VM setup. should run fine in virtual.

KennyG
Oct 22, 2002
Here to blow my own horn.

Corvettefisher posted:

Well that depends, there really isn't a host:vm ratio, look at what the VM's need and address the hosts accordingly.

CF strikes again.

I realize you are normally helpful, but you have a reputation for machine gunning half-rear end'd answers too.

I was not asking about a host vm ratio. I was talking about virtualizing vCenter for 4 hosts running about 40 VMs. If it's ok to run 400+ hosts and potentially tens of thousands of guests on virtualized vCenter than it should work fine for 4/40.

KennyG
Oct 22, 2002
Here to blow my own horn.
Can you boot ESXi from a SAN. Seems there was a fair bit of handwringing about it back on 3.x and 4.x but 5 seems to be supported. Anyone doing it, is it finicky?

KennyG
Oct 22, 2002
Here to blow my own horn.
:emo: I'm looking for a remote VDI solution that is secure but highly accessible... I.e. works over HTTPS, through proxies, hard to block.

Legit business case: workers go on site, need to reach back to the main office desktop to demo the app (not portable). This usually works great over just Remote Desktop (RDWeb) the issue is all the :tinfoil: that network people do to keep their networks safe from the Chineese/NSA.


As I have said we are currently using RDWeb/RDGateway/Hyper-V for this.

We have vSphere/vCenter for another item so Horizon is on the table.

I have requested a sales call from Citrix.

Really, I'm looking for https://myhypervfarm.logmein.com | https://myvdi.join.me - (except hosted on our domain login.mybiz.com/) Please advise of anything worth looking at that is not VMWare or Citrix (is there a separate Micrsoft product I am missing?). Is there a specific aspect of either of those two that is better for what I am looking at. Note it's 12:00am and I have been working these issues for the last 16 hours, so any option is worth consideration if it's more reliable than what we have. I am throwing myself on the mercy of goon wisdom at this point.

KennyG
Oct 22, 2002
Here to blow my own horn.

skipdogg posted:

What's not working? RD Gateway should only need 443 incoming, which everyone should allow out.

You'd think that, but some firms we visit have a weird layered proxy that is killing the connection.

It's bat-poo poo crazy and I am fighting with our management and theirs showing them every step of the way to anyone who will listen that this is their ridiculous config doing stupid :tinfoil: things. Our management wants to throw money at the problem (and bill it back to the client) and while it's not a view I share personally, it's certainly a behavior that, on a Pavlovian basis, I want to reward.


Dilbert, I lost plat on my last ban. Can you email me kj at stillabower dot com

KennyG fucked around with this message at 12:21 on Sep 19, 2013

KennyG
Oct 22, 2002
Here to blow my own horn.

echo465 posted:

Skip the client's internet connection and use an aircard?

I appreciate the brain storming, but it's multiple users at a site, many of whom are the clients. Sometimes we don't even go to the site and it's them (the custoemer) logging in remotely. air cards are just a no go.

KennyG
Oct 22, 2002
Here to blow my own horn.

NevergirlsOFFICIAL posted:

vmware has the worst web site in the world.

sign in to My Vmware then go to downloads-->vsphere, and register

I think my issue is that the marketing is run by some add crackhead. They change the names to products so fast that about the time you figure out what's going on you are completely lost because the poo poo has changed names/been discontinued/replaced/superceeded etc. Why must we rename view to Horizon?

KennyG
Oct 22, 2002
Here to blow my own horn.
VSphere question:

I have 2 vsphere hosts with 2 x E5-2650s and 64GB of ram each.


I have been given the approval to buy (badly needed) more hardware. I want to do two things. One, I have an application that needs pure MHz. I want to get E5-2643 v2 (6@3.5ghz). I also want to add more regular hosts for our general vSphere pool. These I would prefer to upgrade to the new v2 Ivy Bridges as well. Is there a penalty for vMotion or anything besides FT (which I don't want to do) if I don't have 100% identical hardware?

I can separate off vsphere/vcenter clusters if necessary as the application that it is hosting is 100% justified for the demand but it would not be ideal.

Basically, how homogeneous does vSphere need to be?

KennyG
Oct 22, 2002
Here to blow my own horn.

Nukelear v.2 posted:

We've used the Powerconnect 8100 series for a year doing iscsi for our production environment and so far, knock on wood, no issues with it. Is it as cheap as a Netgear, no, but no, you don't want none of that.

Super Secret Saver Pro tip, get the F model and buy the twinax cables as you need capacity, you'll save money and gain more flexibility.

We have the same and have the same remarks.

Stupid easy - plug and play. If you don't get cute, it's rock solid.

Other pro tip, don't buy your cables from Dell. CablesOnDemand.com will save you 70+% Which at 24x3m cables for us was ~5000 vs 1250 - double that for 48 ports obviously

KennyG fucked around with this message at 19:14 on Nov 4, 2013

KennyG
Oct 22, 2002
Here to blow my own horn.
Stupid question time.





Configuration Issues posted:

This host currently has no management network redundancy

What am I missing?

I would think that having two nic's on the management network gives me... management network redundancy. I have tried changing one to standby and one to active but it doesn't work.

KennyG fucked around with this message at 21:23 on Jan 23, 2014

KennyG
Oct 22, 2002
Here to blow my own horn.

KS posted:

Right click on the host and reconfigure for HA -- it will probably clear it.

YAY! Thank you kind sir.

Adbot
ADBOT LOVES YOU

KennyG
Oct 22, 2002
Here to blow my own horn.
Can anyone help me with:

VMHost posted:

Configuration Issues
No vmkcore disk partition is available and no network coredump server has been configured. Host core dumps cannot be saved.

I have 6 hosts in a cluster and 2 of the 6 display the message. The other 4 don't. I didn't do anything fancy to configure them. Can anyone explain what is going on here or how to fix it. I don't think the google diagnosis is what I should do as it's a rather involved SSH/CLI solution that sends me down a rabbit hole I didn't do for the other 4.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply