Time
2 hours 58 minutes
Difficulty
Beginner
CEU/CPE
3

Video Transcription

00:00
Hi,
00:01
you're watching Model fire. Peaceful advanced features Bison three v Sphere, Distributed Resource Scheduler or DRS.
00:10
By the end of this video, you'll understand what DRS does and how to configure it.
00:16
Fees for DRS is another cluster future office fear
00:20
it balances resource consumption across posts by migrating v EMS from overwork hosts to less overworked ones.
00:29
Do your issues is demotion to achieve this.
00:32
So in order to use DRS, you'll need to satisfy all the requirements for the emotion.
00:38
So the arrest is actually an automated be emotion and has three automation levels
00:45
manual where you run DRS manually and choose if you want to migrate some PM's or not
00:52
partially automated where DRS makes migration suggestions to you, but it doesn't apply them until you approve them
01:00
and fully automated, where DRS works everything on itself behind the scenes and present you with the balanced data center Pretty cool.
01:11
In addition to balancing computing load across hosts, DRS can also let you define affinity rules or anti affinity rules for V EMS and hosts. What do I mean by that? Well, suppose you have to virtual machines that should not run on the same years ex I host.
01:29
You can define an anti affinity rule between those two PM's in such a way that when DRS attempts to migrate one of them,
01:37
it takes into consideration the empty affinity you configured and put sitting a different host.
01:44
If the trivia mes are already on the same post, then DRS will detect that and will migrate one of thumb toe another host.
01:51
On the other hand, if you want to vm to run on a specific host
01:56
and not on any other host, you can define an affinity rule between the V M and the hosting question, and you're s will also take that into consideration.
02:06
Years does a lot more interesting things, but for the sake of simplicity, we can't cover them all. That's now jumping to the center and configure DRS.
02:16
Now all rpm's are running on the same host. Remember, we shut down his ex? I too, in the previous lesson and V's fear h a restarted thumb on host one. Now yes, X, I want is running all the V EMS and yes, excite, too, is running. None seems like an unbalanced data center here.
02:35
But before enable and testy rs,
02:38
we need to make sure the VM czar only shared data store
02:42
because DRS. And thus the emotion required the PM's to be only shared data store. The work Virtual Machine is on the local data store, so let's migrated to a shared one using story of emotion.
02:58
Uh, let's do the same with the email VM.
03:07
All right,
03:08
The two database via MES are already on a shared data store. We did that in the previous lesson.
03:15
Now all virtual machines are only shared data store. Before we actually enabled the RS, let us cause a CPU contention in the first years Excite House by stressing the CP use off some virtual machines running on it.
03:30
Because a CPU spike there is a neat utility in Lenox called stress. It lets you strain the CPU and cause it to use all the available computing resources it has. This way we are going to create a clear unbalance in the CPI your consumption between the two hosts.
03:49
So let's do that.
03:50
Opening counsel to the Web VM
03:53
and typing the Command Express Dash C one. The one here represents the number of CPU course you want to stress.
04:01
Okay,
04:03
let's do the same with the email be in.
04:06
It will take some time to settle at 100% of CPO consumption. You can monitor CPU usage on both house and you can see here that yes, X I won is using almost all the computing capacity off its CPU.
04:21
While yes, excited, too, is sitting idle. Since no VM is running one,
04:28
this is most inconvenient.
04:30
We need to fix that by enabling DRS on the cluster.
04:34
So right, click the cluster and choose settings.
04:38
Then click on these few DRS
04:42
and you can see it is turned off. That's turn it on. Click on the edit button,
04:48
then click on the toggle button next to restrict the arrest to enable it
04:53
leave the automation level to the default. It's on fully automated
04:59
and also leave the migration threshold or the aggressiveness level toe default and hit. OK,
05:05
now, if you click on the cluster in the inventory and then go to the summary,
05:11
you can see in the middle pain that there is a tab showing the level of balance in the data center.
05:17
If you refresh the screen, it should show as imbalanced because he is excited. One is using way more CPU than he is. Excited, too.
05:29
Dearest will detect this imbalance and will select one or more vehemence to migrate to host to in order to balance the load.
05:38
So refresh the page and monitor what is about to happen in the recent tusks.
05:44
It could take a dozen seconds to a minute,
05:47
and lo and behold, there it is migrating the Web BM that is causing the trouble to the second years. Excite us.
05:57
Let's wait for the migration to finish,
06:00
and when it's done, you can see that the Web Veum is now running on. Yes, excite, too.
06:06
And if you check the CPU usage for both hosts, you'll see that things look better now
06:14
by migrating the Web PM, DRS. Relieved the first host from some load.
06:18
You can also monitor the cluster balance by clicking on the cluster,
06:24
then on monitor on on CPU utilization. Under these fear, DRS.
06:30
You can see that Host One is using slightly more CPU than Hearst, too.
06:34
But it is way better than before.
06:38
It is difficult to find the perfect balance when you only have two hosts, but things are already looking way better. Thanks to these were DRS.
06:48
That was a quick demonstration of what piece for the arrest does.
06:54
Thanks for watching. And I'll see you in the next video.

Up Next

VMware Foundations

In the VMware Foundations course, students will learn virtualization basics and core concepts, the components and features of the VMware vSphere, and managing virtual machines. Upon completion, students will have a basic understanding of the vSphere.

Instructed By

Instructor Profile Image
Yacine Benbelkacem
IT Trainer and Consultant at CISCO, VMWare
Instructor