Tip:
Highlight text to annotate it
X
00:00:00 - Well everyone, in this Nugget we are going to set ourselves
00:00:04 - up for the future. 00:00:06 - What do I mean by this?
00:00:07 - I want to give you a good solid foundation in UCS
00:00:12 - Manager pools, policies, 00:00:16 - templates, and service profiles.
00:00:19 - These are the ingredients that data center administrators use
00:00:23 - in the UCS to be able to instantly, at the snap of a
00:00:28 - finger, at the click of a mouse, deploy virtualized
00:00:33 - servers in a real high-availability fashion.
00:00:38 - Let's get started. 00:00:38 - Let me explain this to you.
00:00:40 - And it is definitely a preview of what you're going to get
00:00:44 - into in depth in more advanced training classes here at CBT
00:00:49 - Nuggets, like the classes that you'll find
00:00:51 - in CCNP Data Center. 00:00:54 - Now, specifically in this Nugget
I want to start out 00:00:57 - with you discussing hardware abstraction.
00:00:59 - It's critical for you to understand our objective and
00:01:03 - why we want to do hardware abstraction, so these
00:01:07 - ingredients down here will actually make sense to you.
00:01:11 - We'll talk about how we're going to use resource pools in
00:01:15 - order to make this hardware abstraction simple.
00:01:18 - We'll talk to you about how we'll use profiles and
00:01:21 - template to very quickly deploy the particular data
00:01:25 - center that our organization needs. 00:01:28 - And finally, I'll talk to you about
the power of policies to 00:01:32 - literally dictate how things will
operate 00:01:35 - in your data center.
00:01:36 - Look, I absolutely can't wait. 00:01:38 - We just need to jump in here.
00:01:39 - Let's go for it. 00:01:41 - Now, let's talk about hardware
abstraction, as I promised. 00:01:44 - Let's say we've got this UCS chassis
here. 00:01:48 - And on this particular blade right
here we have a virtual 00:01:53 - machine running.
00:01:55 - And that virtual machine is doing Windows Server 2012 for
00:02:00 - us in SharePoint. 00:02:01 - That's its job, and it's doing
a great job. 00:02:04 - But you know what?
00:02:05 - We need to take down this blade. 00:02:08 - There is something that we need
to engage in this blade 00:02:11 - for maintenance reasons.
00:02:13 - Maybe it's to take the blade out and expand the amount of
00:02:16 - RAM that is installed in that particular blade.
00:02:20 - Now, if you've been around the block for a little bit or
00:02:22 - taken some of my courseware in virtualization, you know that
00:02:26 - VMware allows us to do something pretty remarkable.
00:02:30 - We can vMotion that virtual machine to another
00:02:36 - blade, can't we? 00:02:37 - Sure, we can literally in software--
00:02:40 - couple clicks of the mouse-- 00:02:41 - move that VM to another functional
blade. 00:02:45 - This is rather miraculous, and
this is one of the reasons we 00:02:49 - have such high availability when
it comes to a virtualized 00:02:53 - environments.
00:02:54 - Heck, it's one of the reasons that we want to virtualize to
00:02:58 - begin with, the ability to do this. 00:03:01 - But wait a minute.
00:03:02 - When that particular server was living there it needs a
00:03:07 - MAC address. 00:03:09 - When that server was living there,
if it's connecting to 00:03:12 - the SAN, it needs worldwide names
that we discussed, 00:03:16 - right, that are used in Fibre Channel
environments. 00:03:20 - These things are traditionally
burned in. 00:03:24 - These things are traditionally
unflexible. 00:03:28 - Well, in our Unified Computing
world from Cisco we're going 00:03:33 - to virtualize this stuff.
00:03:37 - We're going to make it so that worldwide node names and MAC
00:03:41 - addresses are not tied to the physical hardware anymore.
00:03:47 - That's what we mean by hardware abstraction.
00:03:51 - If we don't tie these names that traditionally were tied
00:03:55 - to the hardware, if we no longer do that, then we'll
00:03:58 - have the flexibility to move something like this virtual
00:04:02 - machine around and not have any issues in
00:04:07 - that particular process. 00:04:09 - You know what we call this?
00:04:11 - We call this stateless computing. 00:04:16 - And it is absolutely the rage,
right? 00:04:20 - Yeah, we want all of these identities
to be 00:04:24 - administratable now.
00:04:28 - And we want them to be controlled by us, the
00:04:31 - administrator, no longer tied to the hardware.
00:04:35 - And this is referred to, this revolution is referred to as
00:04:40 - stateless computing. 00:04:42 - It's hardware identifiers and things
like firmware versions 00:04:47 - being abstracted into, in the UCS
terminology, what we call 00:04:54 - a service profile.
00:04:57 - So we have these identities that were once burned in in a
00:05:02 - software construct called our service profile.
00:05:07 - Each compute node in our Unified Computing System is
00:05:12 - going to receive a unique service profile.
00:05:19 - And what that service profile can do for us is it can
00:05:23 - provide to that particular server, it can provide
00:05:28 - identity like the MAC and the worldwide node names.
00:05:34 - It can provide policy for things like what version of
00:05:42 - the firmware are we going to utilize? 00:05:44 - What are the QoS settings?
00:05:46 - And finally, this service profile can define the LAN or
00:05:52 - the SAN connectivity. 00:05:54 - What is going to be the state of
the virtual network 00:05:59 - interface card?
00:06:00 - What is going to be the state of the
00:06:02 - virtual host bus adapter? 00:06:05 - All very, very powerful constructs
here, powerful, 00:06:10 - revolutionary methods to actually
engage in 00:06:15 - server-based computing.
00:06:17 - So a demonstration is worth a thousand words, right?
00:06:21 - Let's jump into our UCS Manager, and let me show you
00:06:26 - these software constructs that are available to us to make
00:06:30 - this hardware abstraction a reality. 00:06:34 - Let's jump into UCS Manager and
see this. 00:06:38 - So here we are in the UCS Manager.
00:06:40 - And you're going to want to provision for your data center
00:06:45 - an environment where when something needs a name it can
00:06:50 - grab it from a pool that they UCS manages.
00:06:55 - How cool. 00:06:56 - One of the names that servers need
is a 00:06:59 - unique identifier, right?
00:07:02 - This is usually in the BIOS of the particular server.
00:07:06 - But no longer. 00:07:07 - We want to abstract that.
00:07:09 - So if we were to go over to the Servers tab, sure enough
00:07:12 - one of the constructs that we have is Pools.
00:07:16 - And if you go under here, under UUID Suffix Pools, we
00:07:22 - have a pool default that's been created.
00:07:24 - We can assign a pool of unique identifiers, universally
00:07:31 - unique identifiers for servers in our environment.
00:07:36 - So when we're implementing a new server and we want to give
00:07:39 - it a new universally unique identifier, we don't want to
00:07:43 - rely on what's in its BIOS. 00:07:45 - We want to abstract that.
00:07:48 - We want to make that a software-driven and managed
00:07:51 - thing by the UCS. 00:07:53 - We can create a pool of UUIDs to
do that. 00:07:59 - Look at this, we can even gather
up servers into pools. 00:08:04 - Here I have a server pool called
Blade Pool 2. 00:08:08 - And I can take blade servers and
I can put it in this 00:08:12 - particular pool.
00:08:14 - Now, when we want a physical hardware resource we can have
00:08:19 - the UCS Manager do that for us. 00:08:22 - We can automate that process.
00:08:24 - We can say, hey, give us a server from one of pools that
00:08:28 - we've created of servers to help automate deployment in
00:08:34 - the data center. 00:08:35 - This is just so amazingly powerful.
00:08:37 - And I've got a big smile on my face, because I'm recalling
00:08:41 - the first time I saw this. 00:08:43 - For some of you, it's the first
time you're seeing it, 00:08:45 - and I remember how just absolutely
blown away I was at 00:08:50 - the level of power that we are
dealing with here. 00:08:55 - So as we move through the various
tabs in our UCS 00:09:03 - Manager, you're going to have the
ability to automate the 00:09:09 - functioning of your data center
00:09:11 - provisioning through pools. 00:09:13 - Let me show you one more example.
00:09:15 - If you go to the LAN tab, sure enough there is a
00:09:19 - Pools node in here. 00:09:21 - And what do you think would be
in there? 00:09:24 - Of course, your ability to create
a MAC address pool. 00:09:30 - Now when anything needs a MAC address
in the LAN 00:09:33 - environment, it can pull from a
pool of MAC addresses that 00:09:39 - you have set up in this
00:09:41 - unbelievably powerful software. 00:09:45 - And by the way, some pools can
be dynamically created. 00:09:50 - For instance, server pools, you
can go in and you can 00:09:56 - create a new server pool.
00:09:59 - And when you create that particular server pool you can
00:10:04 - literally say, you know what, I want a device to go into
00:10:09 - that pool if it has certain criteria. 00:10:14 - So if a server has this amount
of RAM and that amount of CPU, 00:10:19 - then dynamically put it into the
particular pool. 00:10:24 - So this adds even more power to
the pooling concept, the 00:10:28 - thought of devices automatically
being pooled 00:10:33 - together based on certain features
that 00:10:37 - those devices contain.
00:10:39 - So I think at this stage we really understand
00:10:41 - the concept of pools. 00:10:44 - Now let's take a look at policies.
00:10:46 - So I'll go over to the Server tab here, and I'll go to
00:10:49 - Policies, and I'll expand this node.
00:10:52 - And look at this BIOS defaults and BIOS policies, host
00:10:58 - firmware packages, scrub policies, all these policies
00:11:03 - that we can enact. 00:11:05 - What are these particular things
doing? 00:11:09 - Well sure enough, they are dictating
how the particular 00:11:13 - servers we deploy are going to
react. 00:11:18 - For instance, boot policies would
literally allow us to 00:11:23 - control the order that the server
will try and boot. 00:11:27 - Will it try and boot from the local
disk first? 00:11:31 - Will it try and boot from the CD-ROM
first? 00:11:34 - Will it try and boot from a floppy
disk? 00:11:36 - Whoa, we're going way back in time
there. 00:11:39 - So policies are going to dictate
how a particular 00:11:43 - server behaves within the UCS system.
00:11:47 - And you know the construct that we use to
00:11:50 - bring it all together? 00:11:52 - It's called a service profile.
00:11:55 - Here in the Servers tab I'll go to the Service Profile
00:11:58 - node, and I'll right click, and I'll say let's create a
00:12:03 - service profile, and we want the expert walkthrough here.
00:12:09 - So look at this. 00:12:10 - It says, all right, name the service
profile. 00:12:13 - I'll just say this is our test
profile. 00:12:16 - And by the way, it needs to be
probably following certain 00:12:23 - character constructs like no spaces.
00:12:25 - So there we go, test_profile. 00:12:28 - And it says, all right, where do
you want to get universally 00:12:32 - unique identifiers from?
00:12:35 - And I say, OK, well we've got a pool that
00:12:37 - we created for that. 00:12:39 - So I'll just grab the default pool.
00:12:40 - Notice if we had forgotten to do this we can go ahead and
00:12:44 - create a suffix pool right from here.
00:12:48 - Then we go onto networking, and we define how we are going
00:12:53 - to connect to the network. 00:12:56 - Is it going to be with virtual
NICs? 00:12:58 - Is it going to be with virtual
host bus adapters, right? 00:13:04 - Actually, that would be under the
storage. 00:13:05 - So for the networking it's LAN
connectivity. 00:13:08 - For the storage it's our SAN connectivity.
00:13:12 - Then zoning, then the server boot order.
00:13:16 - So we can pick a particular policy that will be utilized
00:13:21 - for boot order. 00:13:23 - You see what we're doing.
00:13:26 - We are taking a service policy and we're pulling together the
00:13:32 - elements of our pools and our policies that we've defined,
00:13:38 - so that when we deploy a server we associate it to this
00:13:43 - service policy, and all those necessary ingredients are
00:13:49 - pre-defined. 00:13:50 - Yeah.
00:13:51 - Our boot order, the way in which we're going to connect
00:13:55 - to the LAN and the SAN, the way in which we're going to
00:13:58 - assign names, like MAC addresses, worldwide names,
00:14:04 - universally unique identifiers, all of that is
00:14:07 - defined in the service profile. 00:14:11 - And guess what?
00:14:13 - Obviously, we can create service profile templates so
00:14:18 - that we have a standard set of settings that we can quickly
00:14:23 - copy and tweak just a little bit as we need to in a
00:14:28 - particular area of your data center. 00:14:32 - So in this Nugget we didn't really
focus on certification. 00:14:37 - Instead we went a bit beyond certification,
and we built a 00:14:42 - foundation for more advanced courses
to come. 00:14:47 - We discussed how hardware abstraction
is critical in a 00:14:52 - virtualized data center.
00:14:54 - We want to be able to take servers that are virtual
00:14:59 - machines and move them around in the infrastructure and not
00:15:03 - have any problems when it comes to hardware-based
00:15:07 - addressing, like MAC addresses and worldwide names.
00:15:12 - We saw in the UCS how we can create pools of these names,
00:15:17 - how we have policies that dictate how the servers will
00:15:20 - react, and how we can tie these together
00:15:24 - into service profiles. 00:15:27 - We can even create templates of
service profiles to 00:15:32 - simplify the management of those
service profiles. 00:15:38 - Again, a nice foundation for advanced
utilization of the 00:15:44 - UCS system that we will explore
in depth in further 00:15:48 - training here at cbtnuggets.com
00:15:51 - I hope this Nugget has been informative for you, and I'd
00:15:55 - like to thank you for viewing.