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
Red Robin Hood
Jun 24, 2008


Buglord

evil_bunnY posted:

This is not how a modern AD works. Go ask in the thread Docjowles linked.

It is 2003, so not modern. Thank you for the quick link I'll check there!

Adbot
ADBOT LOVES YOU

ragzilla
Sep 9, 2005
don't ask me, i only work here


Red Robin Hood posted:

It is 2003, so not modern. Thank you for the quick link I'll check there!

It's not NT4 with the PDC/BDC design, it's modern.

sanchez
Feb 26, 2003
The PDC/BDC thing is funny, people who were still in school when 2003 came out use the term as though it means something. I blame naming conventions on servers.

evil_bunnY
Apr 2, 2003

Also PDC emulators.

Red Robin Hood
Jun 24, 2008


Buglord

ragzilla posted:

It's not NT4 with the PDC/BDC design, it's modern.

I realized that after the fact. I really do appreciate the clarification.

mAlfunkti0n
May 19, 2004
Fallen Rib
Today I learned that if you are migrating storage from one storage server to another and LUN ID's change, re-signature the disks otherwise you're not going to like the results. I have a feeling my only choice is to present new storage and svmotion my VMs to it and re-signature the disks.

Currently they are seen as snapshots since the LUN id's no longer match the signature on disk. Ughhh.

adorai
Nov 2, 2002

10/27/04 Never forget
Grimey Drawer

mAlfunkti0n posted:

Currently they are seen as snapshots since the LUN id's no longer match the signature on disk. Ughhh.
I do this all the time and have literally never had that problem.

mAlfunkti0n
May 19, 2004
Fallen Rib

adorai posted:

I do this all the time and have literally never had that problem.

I swear in our tests of this exact procedure we had no problems like this. I was sitting there today trying to figure out why one of my hosts was just running at peak capacity and the other two had almost nothing. Then I noticed that they didn't have the storage groups mounted. A dive into the vmkernel log provided me with the "oh hey it's a snapshot we don't like it" view.

Internet Explorer
Jun 1, 2005





Yeah, that doesn't sound right. What storage?

mAlfunkti0n
May 19, 2004
Fallen Rib

Internet Explorer posted:

Yeah, that doesn't sound right. What storage?

It's all FC storage presented through several Falconstor NSS servers. I think the backend is EMC Symmetrix, but it is virtualized by the NSS product.

Vulture Culture
Jul 14, 2003

I was never enjoying it. I only eat it for the nutrients.
This is really common and well-documented behavior when you do a block-level copy of a VMFS datastore. When VMFS metadata is written onto a volume, one of the pieces of information that's written for recordkeeping is the NAA ID of the LUN that the datastore was created on. (In the olden days before NAA support was common, it stored the LUN number instead.)

When you mount a datastore, the VMFS driver checks that the NAA ID written in the LUN metadata matches the actual NAA ID of the LUN hosting that filesystem. If they don't match, ESXi makes the (usually correct) assumption that it's looking at a volume clone or a snapshot that's been presented back to the host so the host can copy files off of it, and it should be treated as a read-only copy of the datastore.

Resignaturing is just a way of rewriting that NAA ID on disk so it matches the NAA ID of your new LUN. The only disadvantage of resignaturing is that it also causes vSphere to generate a new GUID for the datastore, which means that you'll lose all the references to the VMs you have running off of that datastore.

Ubiquitous Talk posted a great article on the underlying structure of VMFS metadata a few years back:
http://blog.laspina.ca/ubiquitous/understanding-vmfs-volumes

Moey
Oct 22, 2010

I LIKE TO MOVE IT
Someone care to assist me in making sure I am reading esxtop right?

Below is a picture of a host cpu tab sorted by %RDY (top entry being idle). I have read that below 10% is showing little CPU contention. I assume I would see a value of 10.00 or something greater if that were the case?

Internet Explorer
Jun 1, 2005





Misogynist posted:

This is really common and well-documented behavior when you do a block-level copy of a VMFS datastore. When VMFS metadata is written onto a volume, one of the pieces of information that's written for recordkeeping is the NAA ID of the LUN that the datastore was created on. (In the olden days before NAA support was common, it stored the LUN number instead.)

When you mount a datastore, the VMFS driver checks that the NAA ID written in the LUN metadata matches the actual NAA ID of the LUN hosting that filesystem. If they don't match, ESXi makes the (usually correct) assumption that it's looking at a volume clone or a snapshot that's been presented back to the host so the host can copy files off of it, and it should be treated as a read-only copy of the datastore.

Resignaturing is just a way of rewriting that NAA ID on disk so it matches the NAA ID of your new LUN. The only disadvantage of resignaturing is that it also causes vSphere to generate a new GUID for the datastore, which means that you'll lose all the references to the VMs you have running off of that datastore.

Ubiquitous Talk posted a great article on the underlying structure of VMFS metadata a few years back:
http://blog.laspina.ca/ubiquitous/understanding-vmfs-volumes

Interesting. Thanks for the insight. I haven't done a block level copy of a VMFS datastore before. Always done a svmotion. Was thinking about the LUNs I've done that with in Windows and haven't had a problem.

1000101
May 14, 2003

BIRTHDAY BIRTHDAY BIRTHDAY BIRTHDAY BIRTHDAY BIRTHDAY FRUITCAKE!

Moey posted:

Someone care to assist me in making sure I am reading esxtop right?

Below is a picture of a host cpu tab sorted by %RDY (top entry being idle). I have read that below 10% is showing little CPU contention. I assume I would see a value of 10.00 or something greater if that were the case?



Thats a healthy %RDY column in esxtop yes. Generally anything over 1% means you're starting to run into contention. From there it becomes "how much can my applications/users tolerate before my phone starts to ring."

When you're under 1% you've got headroom to go!

Moey
Oct 22, 2010

I LIKE TO MOVE IT

1000101 posted:

Thats a healthy %RDY column in esxtop yes. Generally anything over 1% means you're starting to run into contention. From there it becomes "how much can my applications/users tolerate before my phone starts to ring."

When you're under 1% you've got headroom to go!

That's what I thought, just wanted a second brain to tell me that .15 doesn't actually mean 15%.

Thanks!

Also posting a link to this esxtop cheatsheet that I have fallen in love with.

http://www.vmworld.net/?page_id=1156

FISHMANPET
Mar 3, 2007

Sweet 'N Sour
Can't
Melt
Steel Beams
I want to deploy Server 2012 onto an ESXi 5.1 system using SCCM. So first step is I need the network drivers, because I'm using vmnext3. All the guides say I should grab the drivers out of C:\Program Files\VMware Tools\Drivers from a client that has the tools installed, but the clients I'm looking at don't have that driver. I've looked at a 2008 R2 machine and a 2012 and neither have this drivers folder. I've looked on the VMware tools install disk and the only driver on that disk is for the SCSI driver.

So where can I find the mythical vmnext.inf?

Dilbert As FUCK
Sep 8, 2007

by Cowcaster
Pillbug
You should be able to find it when you mount the vmtools install into a VM.

Or you can do it the cheap way and go to a VM that is using the VMXNET, open device manager, right click on the net adapter, driver details, then follow the path to the *.inf and copy it to a share or something.

Moey posted:

That's what I thought, just wanted a second brain to tell me that .15 doesn't actually mean 15%.

Thanks!

Also posting a link to this esxtop cheatsheet that I have fallen in love with.

http://www.vmworld.net/?page_id=1156

Yeah if your VM's were at 15% ready your users would already be calling you asking what the gently caress is going on. What you are seeing is good,

Dilbert As FUCK fucked around with this message at 01:45 on Apr 10, 2013

mAlfunkti0n
May 19, 2004
Fallen Rib

Corvettefisher posted:

You should be able to find it when you mount the vmtools install into a VM.

Or you can do it the cheap way and go to a VM that is using the VMXNET, open device manager, right click on the net adapter, driver details, then follow the path to the *.inf and copy it to a share or something.


Yeah if your VM's were at 15% ready your users would already be calling you asking what the gently caress is going on. What you are seeing is good,

We have a very active environment and if you were seeing over 15% ready values on individual CPUs you indeed would have angry users (been there, done that, over and over). High ready time will also manifest itself as high CPU utilization in task manager (or any other CPU usage graph). We have been trying to train users who request VMs not to shoot for the moon with the size of their VMs. Many of them feel that they need above 1 or 2 CPUs when the reality is they just do not. However they still request it and they get terrible performance from them. Always fun to watch their jaw drop when you lower the CPU count and performance *magically* improves.

mAlfunkti0n
May 19, 2004
Fallen Rib

FISHMANPET posted:

I want to deploy Server 2012 onto an ESXi 5.1 system using SCCM. So first step is I need the network drivers, because I'm using vmnext3. All the guides say I should grab the drivers out of C:\Program Files\VMware Tools\Drivers from a client that has the tools installed, but the clients I'm looking at don't have that driver. I've looked at a 2008 R2 machine and a 2012 and neither have this drivers folder. I've looked on the VMware tools install disk and the only driver on that disk is for the SCSI driver.

So where can I find the mythical vmnext.inf?

I found it in c:\Program Files\VMware Tools\Drivers\VMXNET3\

If you cant find them I can zip them and put them out on dropbox.

talaena
Aug 30, 2003

Danger Mouse! Power House!
Can anyone point me to a resource for Cloud Director's network aspect? I am simply lost when it comes to Org's and their networks and routing through via vShield Manager.

In a nutshell, I want to deploy an AppDirector ovf (or any ovf, the process should? be the same) from Cloud Diretor and be able to access it on the same subnet. I've built the lab from scratch. I have vCenter on 10.1.1.5, vCD us on 10.1.1.8. I deployed the vShield Manager ovf from vCenter and it sits at 10.1.1.100. I have Cloud Director working to the point where I can deploy the AppDirector VM, pop out a console, and configure it's operational settings. The final step that is completely mystifying me is how to setup Cloud Director's Org network to allow me to ping 10.1.1.x; with x being the external IP for the AppDirector VM.

I assume I would be using a routed network, NATing 10.1.1.x into an internal IP address on the AppDirector VM of 192.168.1.y as an example. That part sort of makes sense. I imagine that's what I want, but I have no clue how to get to that point with Cloud Director. Networking is not my strong suite. I had to deploy an edge device and setup networking within Cloud Director; but it doesn't mean I did it right or understand what to do. I just typed in numbers to let the wizard finish. I'm one of those people...

some kinda jackal
Feb 25, 2003

 
 
Maybe a question out of left field, but do you guys think HP's Moonshot will have any significant effect on the virtualization world? It seems like the hardware is way too weak to do much consolidated virtualization. I mean I'm not sure how much you can throw on an Atom with 8 gigs of ram. I don't have too much info on what else will be available on the Moonshot platform though.

I gather that the point is that you can have like 70 blades per enclosure, so maybe you could have like 70 hypervisors with one or two VMs on each, depending on the application?

And I mean this is totally not a serious IT type question; I just wanted your guys opinion :)

some kinda jackal fucked around with this message at 06:00 on Apr 10, 2013

hackedaccount
Sep 28, 2009
I think their main selling point is power consumption. Less consumption means decreased energy bills from the servers themselves, decreased cooling costs, and less wear-and-tear on the server and hopefully a longer lifespan.

Pile Of Garbage
May 28, 2007



Martytoof posted:

Maybe a question out of left field, but do you guys think HP's Moonshot will have any significant effect on the virtualization world? It seems like the hardware is way too weak to do much consolidated virtualization. I mean I'm not sure how much you can throw on an Atom with 8 gigs of ram. I don't have too much info on what else will be available on the Moonshot platform though.

I gather that the point is that you can have like 70 blades per enclosure, so maybe you could have like 70 hypervisors with one or two VMs on each, depending on the application?

And I mean this is totally not a serious IT type question; I just wanted your guys opinion :)

From having a look at the Moonshot datasheet it sounds like it is geared towards running specialised applications/workloads which require significant levels of parallel processing across multiple nodes.

Each "ProLiant Moonshot Server" (Basically a blade for the Moonshot 1500 chassis) only has a single Intel Atom 2.0GHz CPU (With a tiny 1MB cache) and 8GB of RAM (1x8GB DIMM) so I really don't see it as being capable of virtualising anything significant.

Oh and according to the datasheet the Moonshot 1500 chassis takes up 4.3U :wtf:

hackedaccount posted:

I think their main selling point is power consumption. Less consumption means decreased energy bills from the servers themselves, decreased cooling costs, and less wear-and-tear on the server and hopefully a longer lifespan.

Yeah that's definitely the angle that they are pushing. Apparently the 1500 chassis runs off just two 1200W redundant hot-swap PSUs which is gently caress all when compared to something like an IBM BladeCentre E chassis which takes four 2320W redundant hot-swap PSUs.


Jesus, I'm starting to sound like a HP shill (I'm actually an IBM shill :)).

Pile Of Garbage
May 28, 2007



talaena posted:

Can anyone point me to a resource for Cloud Director's network aspect? I am simply lost when it comes to Org's and their networks and routing through via vShield Manager.

In a nutshell, I want to deploy an AppDirector ovf (or any ovf, the process should? be the same) from Cloud Diretor and be able to access it on the same subnet. I've built the lab from scratch. I have vCenter on 10.1.1.5, vCD us on 10.1.1.8. I deployed the vShield Manager ovf from vCenter and it sits at 10.1.1.100. I have Cloud Director working to the point where I can deploy the AppDirector VM, pop out a console, and configure it's operational settings. The final step that is completely mystifying me is how to setup Cloud Director's Org network to allow me to ping 10.1.1.x; with x being the external IP for the AppDirector VM.

I assume I would be using a routed network, NATing 10.1.1.x into an internal IP address on the AppDirector VM of 192.168.1.y as an example. That part sort of makes sense. I imagine that's what I want, but I have no clue how to get to that point with Cloud Director. Networking is not my strong suite. I had to deploy an edge device and setup networking within Cloud Director; but it doesn't mean I did it right or understand what to do. I just typed in numbers to let the wizard finish. I'm one of those people...

When you say "external IP for the AppDirector VM" are you referring to a publicly-routable WAN IP address (E.g. 74.125.237.33)?

some kinda jackal
Feb 25, 2003

 
 

cheese-cube posted:

From having a look at the Moonshot datasheet it sounds like it is geared towards running specialised applications/workloads which require significant levels of parallel processing across multiple nodes.

That's what my understanding is as well, I just didn't know if I was missing some angle :)

madsushi
Apr 19, 2009

Baller.
#essereFerrari
16 under-powered CPUs and 128GB of RAM per U is not high-density, so it's clearly a power play (no hockey pun intended).

If it's 45-blades-per-4.3U, that's an even lower density. - 11 CPUs / 88GB of RAM per U.

madsushi fucked around with this message at 06:52 on Apr 10, 2013

talaena
Aug 30, 2003

Danger Mouse! Power House!

cheese-cube posted:

When you say "external IP for the AppDirector VM" are you referring to a publicly-routable WAN IP address (E.g. 74.125.237.33)?

Not in this case. I was thinking, perhaps incorrectly, about a routed network within CLoud DIrector; where there is an internal IP for the VMs and "external" IPs that route through an edge gateway. That's about as well as I can explain it before my mind goes to mush. Perhaps I'm just thinking about this all wrong. I'm currently fighting with perl on an HP/UX 11i machine in the other window; so my brain is mush.

1000101
May 14, 2003

BIRTHDAY BIRTHDAY BIRTHDAY BIRTHDAY BIRTHDAY BIRTHDAY FRUITCAKE!

talaena posted:

Not in this case. I was thinking, perhaps incorrectly, about a routed network within CLoud DIrector; where there is an internal IP for the VMs and "external" IPs that route through an edge gateway. That's about as well as I can explain it before my mind goes to mush. Perhaps I'm just thinking about this all wrong. I'm currently fighting with perl on an HP/UX 11i machine in the other window; so my brain is mush.

If you want the app director Vm to have an ip on 10.1.1.0/24 then you probably want to create what's called an external network in vcloud. This maps directly to a real vlan/VMware port group.

No reason to NAT in your case.

Tomorrow after I finish doing some training for a customer I can step you through the process if its still a little unclear. Optionally you could just import the ovf directly into the vcenter inventory.

Edit: once you create that external network you need to add it to an org vdc then it should be available to use for virtual machines.

Pile Of Garbage
May 28, 2007



talaena posted:

Not in this case. I was thinking, perhaps incorrectly, about a routed network within CLoud DIrector; where there is an internal IP for the VMs and "external" IPs that route through an edge gateway. That's about as well as I can explain it before my mind goes to mush. Perhaps I'm just thinking about this all wrong. I'm currently fighting with perl on an HP/UX 11i machine in the other window; so my brain is mush.

That's alright. Based on your first post I thought it was just a network configuration question however after reading up on vCloud Director (Which I'm honestly not familiar with) it seems a bit more complicated than that. I guess a good place to start would be with the vCloud Director Administrator's Guide, specifically this section.

Erwin
Feb 17, 2006

Anyone know what causes this poo poo?



It's happened twice. Last time support essentially shrugged saying "iunno" before telling me to use the stand-alone converter to clone the machine. It was an unimportant machine and I don't remember what I actually did, but that's bullshit. This one is important and I'm not using the stand-alone converter to work around some stupid loving bug.

Edit: vSphere 5.1.

Dilbert As FUCK
Sep 8, 2007

by Cowcaster
Pillbug
Err wait I think I read that pic wrong, can you post a pic of the VM folder?

Erwin
Feb 17, 2006

Corvettefisher posted:

Err wait I think I read that pic wrong, can you post a pic of the VM folder?

Well, that'd require a lot of redacting, but there are two child disk files (-000001.vmdk and -000002.vmdk) for each disk. The '000002' files are all 17,408KB, which I think is the minimum size of a child disk? This was a manually-taken snapshot.

Frozen Peach
Aug 25, 2004

Heroes Never Die

Our VM host is connected by fiber to a SAN. Both of our hosts connect to the same LUN as a shared datastore. Is it safe to resize the datastore LUN on the SAN, and then resize the datastore itself on the VM host while both hosts are running and VMs are running on them? Is there any special precaution I need to take?

talaena
Aug 30, 2003

Danger Mouse! Power House!

1000101 posted:

If you want the app director Vm to have an ip on 10.1.1.0/24 then you probably want to create what's called an external network in vcloud. This maps directly to a real vlan/VMware port group.

No reason to NAT in your case.

Tomorrow after I finish doing some training for a customer I can step you through the process if its still a little unclear. Optionally you could just import the ovf directly into the vcenter inventory.

Edit: once you create that external network you need to add it to an org vdc then it should be available to use for virtual machines.

This is exactly what I need and exactly what confuses the gently caress out of me when I get into the vCD interface. I simply don't understand the concepts. I can (and have) read through the bits of the Admin Guide and I think my spleen processes more of the information than my brain; so I'm still stuck drooling like an idiot.

I will have to play more with the external network aspect instead of trying to NAT. I *almost* understand the process, but not really. I just need someone to show it to me, treat me like a 5 year old, and once I see it actually working it might sink in.

Moey
Oct 22, 2010

I LIKE TO MOVE IT

Frozen-Solid posted:

Our VM host is connected by fiber to a SAN. Both of our hosts connect to the same LUN as a shared datastore. Is it safe to resize the datastore LUN on the SAN, and then resize the datastore itself on the VM host while both hosts are running and VMs are running on them? Is there any special precaution I need to take?

None that I know of as long as your SAN supports it. I have done it dozens of times without any performance interruption or outages.

What kinda SAN?

FISHMANPET
Mar 3, 2007

Sweet 'N Sour
Can't
Melt
Steel Beams

FISHMANPET posted:

I want to deploy Server 2012 onto an ESXi 5.1 system using SCCM. So first step is I need the network drivers, because I'm using vmnext3. All the guides say I should grab the drivers out of C:\Program Files\VMware Tools\Drivers from a client that has the tools installed, but the clients I'm looking at don't have that driver. I've looked at a 2008 R2 machine and a 2012 and neither have this drivers folder. I've looked on the VMware tools install disk and the only driver on that disk is for the SCSI driver.

So where can I find the mythical vmnext.inf?

Well, I figured out my problem. On my Server 2012 box at least, the files are in C:/Program Files/Common Files/VMware/Drivers/.

Dilbert As FUCK
Sep 8, 2007

by Cowcaster
Pillbug

Frozen-Solid posted:

Our VM host is connected by fiber to a SAN. Both of our hosts connect to the same LUN as a shared datastore. Is it safe to resize the datastore LUN on the SAN, and then resize the datastore itself on the VM host while both hosts are running and VMs are running on them? Is there any special precaution I need to take?

Nope growing a VMFS volume is a live non impacting of production services.

Erwin posted:

Well, that'd require a lot of redacting, but there are two child disk files (-000001.vmdk and -000002.vmdk) for each disk. The '000002' files are all 17,408KB, which I think is the minimum size of a child disk? This was a manually-taken snapshot.

Not exactly sure here. Is there anything asking you to consolidate?

talaena posted:

This is exactly what I need and exactly what confuses the gently caress out of me when I get into the vCD interface. I simply don't understand the concepts. I can (and have) read through the bits of the Admin Guide and I think my spleen processes more of the information than my brain; so I'm still stuck drooling like an idiot.

I will have to play more with the external network aspect instead of trying to NAT. I *almost* understand the process, but not really. I just need someone to show it to me, treat me like a 5 year old, and once I see it actually working it might sink in.

If reading isn't your thing, there is a free E-Learning course on it here
http://mylearn.vmware.com/mgrreg/courses.cfm?ui=www_edu&a=one&id_subject=43307
http://mylearn.vmware.com/mgrreg/courses.cfm?ui=www_edu&a=one&id_subject=40587

Dilbert As FUCK fucked around with this message at 20:55 on Apr 10, 2013

Frozen Peach
Aug 25, 2004

Heroes Never Die

Moey posted:

None that I know of as long as your SAN supports it. I have done it dozens of times without any performance interruption or outages.

What kinda SAN?

DELL PowerVault

Syano
Jul 13, 2005

Frozen-Solid posted:

DELL PowerVault

All day long dude. Do it and throw your sunglasses on cause youre ballin with shared storage.

Adbot
ADBOT LOVES YOU

talaena
Aug 30, 2003

Danger Mouse! Power House!

Many thanks. Book learnin' is apparently beyond me. Words confuse me. You'd think I'd remember mylearn by now. I'll definitely take both of those courses, will make me feel less like a drooling moron and might actually count towards my 'training' metrics I have to meet; despite the fact I don't support this product.

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