Hello, I'm Dean Pompilio. Welcome to Cyber Eri.
This is Module 11 Lesson one of the virtual ization insulation configuration management course.
So this lesson we're going to discuss high availability.
Look at some of the h A configure options Discuss the responses that h A contain under different circumstances
when a host of virtual machine or an application fails.
So, first of all, how availability gives you various different levels of protection I can protect against hardware failures.
I can plan for maintenance with minimal downtime. We're zero downtime even
and also protect against unplanned downtime Or, in the case of a disaster, if you're
you lose hardware or you lose power for some device than a J can help you maintain availability of your applications for your environment.
I can also work with VM migration
where V EMS could be moved around automatically will cover that a little bit later. In the course,
we can also use multiple io paths.
So if I lose a path to my storage, if I've got a redundant path, I can then have
high availability takeover. In that case,
this is just the generic sense of high availability.
VM load bouncing again with clustered servers,
viens could be adjusted for where they reside in. That's another form of high availability.
And then we also have fault tolerance and disaster recovery.
These are all ways to keep your operations running keeper environment up and available in the event of different kinds of failure scenarios.
availability RV center itself,
some of the best practices that we should follow is to think about clustering the V Center database.
You can do this lots of different ways. It depends on what kind of
third party database server you'd like to use. Whether it's Oracle or sequel server. You have options there.
We should also think about having redundant active directory servers.
You don't want a single point of failure in either one of these two areas
because even if you're the center databases available and you have single sign on configured
if you're active, directory server is down and you can't authenticate with it. Now you're stuck using only local accounts that could break a lot of
configured the V Center VM itself to be a high availability of'em.
That's a great option. So that way If your V Center of virtual machine experiences problems or the host that is running on experiences problems, we can restart it another host and have
availability of the center.
we could also use the helping feature for the center server.
This allows the center to know when there's a problem with the connection to another host or to a data store. And that way it can make decisions on what to do in the event of some kinds of failures.
High availability for V sphere in general
is what we're striving for here.
Minimal downtime. There could be some scenarios where you have some of it, some very small amount of downtime,
but in general everything remains up and running, and the user's usually don't notice any difference.
We also need to remember that this will work with all of the supported guest operating systems and all the support of hardware
that the center is compatible with.
We're protecting against the host failure, VM failure or at failures,
a little bit of information about the your clusters here we could have up to 32 hosts in a cluster, so you can certainly make a very large cluster to absorb multiple failures.
We can also have up to 512 e ems per host.
Of course, you've been a really powerful host to manage 512 E ems on one physical server, but it's possible. And then you can have up to 3000 V EMS per cluster.
So if you had a 16 hosts in your environment, you could build, you know, four quad host clusters and be able to host 12,000 V EMS in that kind of environment. That's a really huge consideration. Most environments are not that large, but maybe yours is,
and you want to be able to know those numbers.
You might see items like that on the exam.
All right, so let's think about some different failure scenarios. First of all, we got a diagram here showing
some storage. Got a cluster with three nodes.
Ah, a couple of'em is running on this host few on that host and one on this host.
if I lose host number one? If it just goes away?
There's a couple of different things that could happen. One thing is that you might end up in what's called host isolation mode.
That means that maybe the connection to the network went away. But the host is still running.
That means the V EMS. Can we still run?
Even though they might not be able to
connect anything until that network connection gets restored? At least they're they're still up and running.
If the host, however, fails
like the power supply dies or it has a memory faulting crashes now, those to be EMS
should automatically pop up on
another host in the cluster.
Most likely, they would go to host three in this scenario so that we're have some level of balancing here.
high availability agent
access a heartbeat, among other things. So it knows the connections between the three hosts are available.
It's checking the health and status of those hosts, and when it detects a problem,
it decides either to move the PM's or keep them where they are, because it might be a problem that's recoverable,
and we'll see a little bit more of those options when we do, the lab,
via more tools, must be installed on the PM's that are running on the hosts in the cluster
for for some of these features to work. So just as I mentioned earlier, we want to always think about having via Mutual's installed on all the EMS.
There's really no good reason not to do it,
unless you just were completely out of storage space and couldn't install the software for that reason.
So if if Of'em fails
or if the heartbeat stops Hervey M, then the H A agent will restart the V M on the same host. So maybe maybe VM five fails. For some reason,
the H agent, which is doing the heartbeat function between the hosts, just restart it again.
So it's pretty resilient to problems of their VM is, of course, if the VM keeps on failing, you should get some kind of notification about this because you maybe you have some alarms set up in your environment like we talked about in the last lesson.
Or you just start to get calls from users complaining that some server is not available.
In the third scenario, if an application fails a J agent again,
we have taps running on Vienna's three and four.
The V M might be restarted on the same host again, so in this scenario,
maybe I've got an APP running on the M six.
This application fails,
so V M six restarts again
and the application should restart with it
so we can see as a recap that weaken configure high availability to protect us against various different problems.
Problems with the host from with the EMS problems with applications,
we know that high availability and a general sense could be used to move the EMS around for low balancing purposes. We can deal with
losing an eye. Oh path
fault tolerance Also possible.
It's a little bit different than high availability. We'll get to that, though.
We talked a little bit about the options for making your V Center virtual machine highly available. And then we talked about the different features of H. A.
We know that it works with all supported guest or less and hardware
and we protect us against hosts. Bm zaps failing.
Then we looked at some of those scenarios.
When the host fails, the VM got started on other hosts of the VM fails of'em get start on the same host
and then if an application fails, the VM gets restarting the same host.
All right, thank you. That concludes lesson one