All language subtitles for 13 - Troubleshooting EC2 Launch Issues English

So let's look why an EC2 Instance,

would have trouble to be launched.

So if you get an Instance Limit Exceeded error,

that means that you have reached

the max number of vCPUs, per region.

So when you launch your EC2 Instance,

it has a specific number of vCPUs.

And there are limits for On-Demand

and Spot instances for vCPU.

For example, if you use standard instances,

then you'll have by default 64 vCPUs available to you.

So the resolution in case you reach that limit

is to launch either the instance in different region

where the limits will be reset,

or to request AWS to increase your limit, in the region.

And just so you note, the vCPU based limits only apply

to running On-Demand and Spot instances.

So we can see this in two ways.

First of all, if you go to Limits on the left hand side

and then we look for vCPU.

We have two limits to look at.

So, the first one is this one,

All Standard Spot Instance Requests,

which is limited to 64 vCPUs.

That means that once you reach that number

then we will have an error.

And if you scroll down,

there's a Running On-Demand All Standard.

So again, it's for 64 vCPUs, okay?

And, in case you wanted to increase that limit,

you could click on here, to request a limit increase.

There's also a way for you to calculate the vCPU limit

using this calculator, but I won't go over it.

And another way to look at it,

would be the Service Quotas console,

which is a very cool service.

So you would go into the Amazon Elastic Compute Cloud,

Amazon EC2, and then you would look for standard,

the keyword standard.

And as you can see, All Standard Spot Instance request

has a 64 limit and Running On-Demand as well,

has 64 limits.

And if I click on it, we can have a look at

what is the value and the quota value,

and then we can request a quota increase.

Now, how do we know how many vCPU get launched

by each instance?

Well, if we go and Launch an instance,

we can find out how many VCPUs, based on Instance type

by going into Instant type,

and then click on Compare instance types.

And in here, you have the Instant type

versus number of VCPUs.

For t2 micros, we can see we have one vCPU,

but if we choose an instance such as the C59 xlarge,

we have 36 vCPU, so using this instance type,

it would quickly reach the limit.

The second kind of troubleshooting issue is

if you get an Insufficient Instance Capacity.

So if you get this error, that means that AWS

has not enough capacity for On-Demand,

in the particular AZ, where the instance is launched.

This is not a problem on you, this is a problem for AWS.

So a resolution for this error can be multiple.

For example, you can wait a few minutes

before requesting again, in case

there is more capacity added to that particular AZ.

Or, you can, for example, request less instances at a time.

For example, say you're requesting 10 instances,

in an AZ, but it's not possible.

Then maybe break it down into smaller requests.

Or, make a request for a different instance type.

For example, something completely different.

And then you can resize later,

as we can see, we can resize our instances

or obviously, please try to launch the request

in a different AZ, if that's something

that is acceptable to you.

And finally, if the instance that you launch

terminates immediately, so it goes

from pending right away to terminated.

That means a couple of things.

You either reach your EBS volume limit, so check this out.

Or your EBS snapshot, you're launching

your Instance from is corrupt,

or the root EBS volume is encrypted

and you do not have permissions

to access the KMS key, for decryption.

Or if it's an AMI, instance store-backed AMI,

then maybe it's missing a part.

So to find the exact reason, you should check out

the EC2 console of AWS, which should give you,

in the description type, the reason why there was

a termination of your instance directly, okay?

Begin to remember, all that I said to you.

So, the Instance terminated, the Instance capacity,

and the Instance limit, going into the exam.

So that's it. I will see you in the next lecture.