2.9 Virtual Networking Part 9

00:00

So with that in mind, what is smart? Paging So

00:04

smart? Paging allows a virtual machine to temporarily use hard disk space to supplement shortages off memory. What we call non volatile memory,

00:18

um, so smart

00:24

during VM restarts when there is no available physical memory. Now what do I mean by this? So

00:30

let's say you have 30 gigs of memory available on your server

00:36

on dhe. 28 gigs is currently being used by all the machines that you're running on that server.

00:43

You have to reboot one of your servers, and it's currently using four gig. But when you start the machine, it means eight. Gig of RAM to stop for whatever reason,

00:53

it now has to find it will look for two extra two gig, which you can say you've got. So you've got a 30 gig server using 28 gig. If you press restart, it will ask for the extra two gig. Fair enough,

01:04

but it now has a surplus. It's asking for more memory than the server physically has available across the whole cluster.

01:11

What smart paging does is the remaining amount of RAM that you are requesting that the VM is requesting

01:19

it creates a file and says, Right, drop all your extra stuff into this file and switching in and out like a swap file like you would find on any other client machine

01:29

and that allows it to load up. What this means is that even when you've got full utilization of your memory, as long as you have space for these swap files and paging files to be made, you can still restart a server, even though you haven't got a Lee available round to do the normal restart

01:46

s. So that's that's what this does so yet to use when memories full data can be transferred in and out of memory by swapping it with other data from the paging file works exactly the same way the normal paging file. But it's doing it on a smarter level because it's only doing it for the amount of memory that you require

02:04

on. This is like I said, it's used in hyper Vito. Allow the EMS to use disk space to supplement a shortage of memory. So this this allows you to

02:14

actually fix the problem of having over utilization of ram.

02:20

As I mentioned, it's only used when you reboot. It's the only time it will not use a paging file for general use like you would in a normal machine. Like Like you are, sadly, your machines right now, or I am. It won't do it then.

02:37

All right, So, um

02:39

yes. So it's when there is no available physical memory on your servant needs to be restarted. It's the only reason to use a smart paging fall. If you have enough, RAM bio means turn this off. He don't need it. But that being said, I always recommend to all my clients you should always leave this on. Just in case

02:58

on why I tend to do is to give an example of a real world environment.

03:01

I tend to have a small partition on the network, which I tell

03:07

if you're gonna make a smart Paige drop it in this partition. So that way, that mission that space is allocated, it's not gonna cause a crash. You will literally get in there a message saying server cannot be restarted. Your VM cannot be restarted if you don't have space to make a smart page file and you've run out of room.

03:23

So it's quite a critical piece. Small, but critical.

03:28

So it's just something to keep in mind. I'm actually going to show you this. The dynamic memory and smart paging itself s o. Let's jump back across the hyper V s a dynamic memory here. So what I'm gonna do is I'm actually going to show you very quickly on here.

03:43

Um, we're gonna go over to the memory you can see here. I've set a two gig amount of ram, and I've told it to drop the minimum round 2 512 It's not being used. And the rest is all the same dynamic memories on. So this is what I've set up. What we're gonna do is I'm gonna start the machine, so I'm going to Yeah, fine. Whatever. Revert that machine.

04:02

That's fine.

04:03

So it's not gonna boot the machine up. Here we go. So you can see here. The state currently on the screen is running on dhe. The assigned memory is 2048 megabytes, so you can see the up time

04:16

and we're currently merging the files. So there are because I haven't used this for a while. This is gonna merge the different files together effectively on dhe. It allows it to basically set itself up, and you can see the up time taking a lot

04:30

down here. You have your checkpoints. This is I'll just explain this. While we're waiting for the item to occur,

04:35

this is where it's taking it from, so it's from this time and date on effectively at the bottom. Here, you can see a summary of the actual thing when it was created, whether it's clustered if you've got a heartbeat, these are more advanced things

04:48

importantly, and you mentioned earlier Lance how you would check if it's a general on a Gentoo without going into settings. If you click on the machine and then come down here, it will actually tell you what generation the machine is as well. So if you just want to quickly check without jumping into different settings and things like that, that's when you would come and you can actually set notes on the machines. Well,

05:09

over here you can see the startup memory, the minimum memory.

05:12

This allows you to see it

05:14

as it's currently working itself. Now you can see

05:16

in the time that we've been speaking, so this machine is now on and running. Okay,

05:21

The startup memory was 2048 Meg. However, now that it's up and running and the machine isn't actually doing much other than sitting on my machine

05:31

because the minimum memories 5 12

05:33

the current assigned memory is 720 Meg and the machine is using 5 60 approximately is going up and down a little bit.

05:43

Why has it got 7 20? Well, that's because that's the memory, but for kicking in. So it's currently using that. You can see it's it's kind of fluctuating a little bit, but it's using 600 Meg, and it goes up and down is different. Updates occur. It's probably going off on checking. It's got the latest Windows updates and service is starting up and all kinds of stuff.

06:00

And it will actually say If you can see here, the memory status is just flicked a low. Which means if this keeps loading up, the hyper visor is going to get a request from the SVM to say, Please, can I have a little bit more round

06:13

on? The more it ticks up and you can see it's kind of slowly climbing, it's gonna get to a point where the hyperdrive is gonna go. Okay? I can see you need a bit more memory. Let me send some more over and it's gonna buffer it up even more

06:25

s so I'm hoping, actually might hit it. It's only like 20 Meg away from hitting the limit on it. It depends because it's waiting for a signal from the V N to say, guys I've run out of Ram. I need some more. Please. Can you give it to me on lo and behold, Bang on time.

06:42

There we go. So it's now given it more memory and it's using even more

06:46

So this is a perfect example of dynamic memory in action. Now, if I was to restart this server,

06:51

let's see if I can do it from here. So I'm not gonna reset because that's literally like pressing the restart button on the machine. What I want to do is shut it down.

07:00

Andi, this is how you would do it in a virtual environment. Anyway, you would say please shut down, set yourself up. And then when I now turn it back on, it's gonna go back up to that start up memory level of 2048 megabytes, so you can see here is currently in the process of shutting down.

07:16

This is one of those If it was a shielded V M and I don't have the access to do this,

07:20

I wouldn't be able to if I wasn't the tenant administrator. Eso this sake is quite nicely into the previous stuff that we've discussed.

07:29

You can see there's still shutting down its waiting for. Basically, it's like you would normally with the machine. It just shuts down stuff. You can see here a little progress window of what's going on on the server. So that's just saying shutting down. And then that would basically just disappear. Server flicks off and this will state will be running off just like this one here

07:46

on, then that's it. There we go. So back off there. If I now hit start again, it's gonna check to see if it's got the Ram. If there's not enough RAM, it will ask for a paging file. So the second I hit start again. There we go is boot in the server back up, and there's your assigned memory of 2048 again. So that's what dynamic memory does. And then smart paging is like

08:05

the support underneath dynamic memory, just in case you run out.

08:09

So that's kind of Ah, quick demonstration off dynamic memory in action. It's very quick to respond, but the hyper visor tries to keep control as much as it can throughout the whole fit,