Lesson 3 Part 1 - Migrating Virtual Machines (VMs)
Video Activity
Lesson 3 part 1. Migrating Virtual Machines (VMs) This lesson is about migrating virtual machines (VMs) and covers the following: Types of migrations The importance of vMotion Identifying and verifying vMotion requirements Perform vMotion migration Performing storage vMotion migration Performing enhanced vMotion migration Migrating a virtual machin...
Join over 3 million cybersecurity professionals advancing their career
Sign up with
Required fields are marked with an *
or
Already have an account? Sign In »
Video Description
Lesson 3 part 1. Migrating Virtual Machines (VMs) This lesson is about migrating virtual machines (VMs) and covers the following: Types of migrations
-
The importance of vMotion
-
Identifying and verifying vMotion requirements
-
Perform vMotion migration
-
Performing storage vMotion migration
-
Performing enhanced vMotion migration
Migrating a virtual machine (VM) means it is being moved from one host of data store to another host or data store. There are several types of migrations: - Cold: VM is powered off
-
Suspended
-
vMotion
-
Storage vMotion
-
Concurrent migrations
Video Transcription
00:04
Hello. I'm Dean Campiglio. Welcome to Cyber ery. Where in this lesson we are in module seven. Lesson three, installation configuration management of virtual ization environment. Now we'll be talking about migrating VM.
00:18
Well, look at the different types of migrations.
00:20
Why? V motion itself is important.
00:24
We'll look at some of the
00:26
requirements,
00:29
see what the migration looks like itself.
00:32
Talk about storage
00:34
of emotion,
00:36
migration.
00:37
And then the last option is the enhanced the motion migration.
00:41
This lesson will also wrap up with lab 11.
00:45
Okay, So migraine of'em means that we're moving it from one host or one day to store to another host or another data store.
00:54
Depending on your requirements and your capabilities, you can either do these things wanted a time or do them both at the same time. So
01:02
pm where does give you a lot of options
01:04
for moving the EMS around
01:07
the different migration types
01:08
are are what we see here. If the VM is powered off, it's considered a cold migration.
01:15
As I mentioned in earlier lecture the cold migration.
01:19
If you can do it if you can get the downtime, is probably a good choice for performance reasons.
01:23
The migration will happen much more quickly if the VM is not up and running
01:29
because there's all that
01:30
processing and memory requirements and network requirements just aren't there when the VM is powered down.
01:36
So think about that option if you if you need to get something done quickly and you're able to,
01:42
uh,
01:42
get the down time
01:44
if you suspend of'em. This is nearly as as effective as the cold migration as faras performance.
01:51
You can hit your pause button suspended. Bm
01:53
The people using it
01:57
basically lose their connection until the VM is booted. Thio
02:00
full state again or un suspended if you will.
02:05
So when it's in that suspended mode, you can you could migrate it
02:07
to get a little bit better performance
02:10
when the PM is powered on. That's what we call an actual demotion
02:15
again between hosts or even with storage the motion. I can now move the files that comprise that VM while the VM is up and running.
02:25
So these are powerful features that let the virtual environment administrator deal with the changing needs of a dynamic environment.
02:34
If you're running out of storage in one place, you can storage the motions and VM somewhere else. If one host is getting overloaded with processor and memory requirements, you could move some of those v EMS to another host that has some available capacity.
02:49
A SZ faras concurrent migrations were able to do eight simultaneous operations here. Whether they are v motion cloning,
02:57
deployment of the EMS or storage of emotions,
03:00
these can all be done with a single B. M. F S five data store.
03:05
So that's a pretty generous allotment of resource is if I could do eight of these at a time. Of course, your performance would greatly be impacted if you do two or three or four of these at once. If you did a full eight,
03:19
it might be possible. But you better have really powerful hardware unless you wanna wait
03:23
for the rest of the afternoon for that to complete cause. It could take quite a long time.
03:30
So let's look at our different migration types and compare some of their
03:32
There are parameters here.
03:36
In the case of the cold migration,
03:38
Power's off. As I mentioned,
03:39
I can migrate
03:42
to a data store or host
03:44
or both at the same time.
03:46
You could do a cold migration from one data center to another,
03:51
and I don't have a shared storage requirement.
03:57
Also, another thing that's nice is I can use hosts that come from different CPU families.
04:03
Ideally, you want your hosts to be as identical as possible.
04:09
Same CPU architecture, same memory,
04:11
same
04:13
basic functionality if you're gonna be purchasing hosts just by two of the same thing.
04:18
But in some cases that's not possible. You might have a legacy system, and you've got some newer hardware.
04:24
So in that case,
04:26
a regular V motion or storage. The motion is not possible if those CPU family's heir different. If you're going from like an A M D system to an Intel system, for instance,
04:34
and those scenarios, you must do a cold the motion in order for it to work.
04:41
That's because
04:42
the since the BM is off once I move it to another host with a different processor architecture. When it boots up,
04:49
it will work within that environment because it's the *** ice offers already running there.
04:56
But if if the VM is running, then the processor families must be compatible.
05:01
So if our V m a suspended the power status suspended,
05:04
just like a cold migration. I could migrate to the host or the day store or both at the same time. In that in that mode
05:13
could also do a suspended migration between data centers, which is handy.
05:16
Also do not need shared storage because the PM's not using the storage at that moment, so it's not a requirement.
05:25
However,
05:26
the CPU compatibility must be in place
05:30
because the V M is
05:31
when you when you un suspend the VM and bring it back to its full running status.
05:36
There is no reboot happening there, so the CPS family must be compatible. In order for that to work
05:44
of emotion,
05:46
which, which requires the VM, are well as possible, I should say with Of'em is on
05:50
I can on Lee do this between hosts that that's what the emotion is for.
05:56
I cannot do this across data centers.
06:00
Shared storage is required. Eso both of the hosts and involved must be accessing the same shared storage.
06:08
And as you would expect, the CPS must be compatible because the BM is up and running
06:13
storage of emotion or I'm moving the files of the V M to another day of store. I could do this while the PM is on.
06:21
I could only do it between data stores.
06:24
Doesn't have all the host
06:26
not possible across data centers.
06:28
No stair shared storage requirement.
06:30
Since I'm just moving the files, I don't need to share anything and then see if your compatibility doesn't apply here, so that's not applicable.
06:40
And then, if you remember, the enhanced the motion is moving the V m from one host to another and from one data store to another at the same time,
06:48
the V M could be on. I can do both the host in the day store,
06:54
not possible across data centers.
06:57
I don't have a shared storage requirement, but again, because the BM is running, I have a CPU compatibility requirement.
07:04
So there's lots of reasons why you would move the EMS around. I mentioned some of these earlier trying to improve your harbor utilization. What we mean that by that is
07:14
that if I've got
07:15
to host or even three hosts
07:17
and the load from V EMS on each of those hoses uneven,
07:24
meaning that I've got very heavy
07:27
CPU and member requirements at host number one ah, moderate amount of host number two in a very low requirement on host number three.
07:33
I should be able to move some of those V EMS around to balance that out.
07:38
So I'm getting better. Use of my available hardware. Resource is by distributing the requirements of the PM's more evenly.
07:45
You can do this manually, of course.
07:47
Um, you can also use DRS, which we'll see later to do this in an automated fashion.
07:54
So DRS will look at your host in the cluster and decide that
07:59
that the memory usage is uneven with CPU uses. It is uneven, so it will make recommendations which could be automatically
08:07
enacted. Or you manually must initiate those recommendations.
08:11
And as we mentioned before, this allows for continuous operation during the V motion process,
08:18
the users might experience a very small performance degradation during the V motion activity.
08:24
But in general, unless you're using an application, has a lot of real time interaction like databases. For instance, most users wouldn't even notice. So the actual steps that happened for of emotion or a rather interesting
08:37
we began by copying the memory state of the BM on the source host to the destination host.
08:46
So since the memory
08:48
information is containing that host. It could be just treated like any other data and shipped across the network
08:52
to the destination host.
08:54
Once that's completed, the VM is qui est,
08:58
and this is the source P. M. One of the source host. And by qui est, we mean that it's basically quieted down to the point where no more memory changes are happening.
09:09
Once the memory changes have stopped on the source host,
09:15
then that V EMS remaining memory can then be copied over the network to complete Step one that these two kind of work a little bit the same time
09:24
Once, once that happens now, the VM could be initialized on the destination host
09:28
and it could be
09:30
I'd start running. It doesn't need to go through the boot process,
09:33
as you would expect.
09:35
It's actually copying the memory contents over, and the machine can start running on the source host.
09:43
Now. Any users that were that were interacting with the PM on the
09:48
I said, I'm sorry, I think I said that wrong. VM is initially from running on the target house, not the source house.
09:54
Any users that were interacting with the BM when it was on the source host can now interact with it on the destination host or the target host.
10:01
They don't know the difference. They don't see any interruption in their connectivity.
10:07
Then on the source host, the memory pages that the VM were used was using can now be freed up and made available for other uses.
10:16
So that's the way
10:18
of
10:20
holding those. Resource is on the source house until they're all the members copied over when everything's finished, then it marks those pages as free and other V EMS can access them.
10:30
All right, stay tuned for a second portion of lesson number three. Thank you.
Up Next
Similar Content
