Route 53 and Networking

Question 1

You want to route all traffic from Europe to EU-WEST1, What routing policy would you use and what location would you specify for your users?

Answer 1

Geolocation and set location to “Europe’. R53 allows you to identify countries or continents

Question 2

What is a multi-value answer routing policy and how do you set one up

Answer 2

An MVA policy will return a list of healthy endpoints to the client (Browser). You create multiple record sets containing an endpoint with associated health checks. When the record set is queried only healthy endpoints are returned and the browser picks one at random.

Question 3

Is the domain registrar the same as a DNS?

Answer 3

No

Question 4

What are the steps to use a 3rd party domain with R53

Answer 4

1. You need to create a hosted zone in R53 using your domain.
2. Update your 3rd party name server records to point to the AWS name servers provided when you create the hosted zone.

Question 5

You have an R53 record-set pointing to an instances IP as the endpoint. You need to change this to an ELB. What record type do you currently have and what is it changed to?

Answer 5

You currently have an A (or AAAA for IPv6) and you will need to change to an Alias

Question 6

There are 4 common record types in R53. What are they and what do they do?

Answer 6

A: URL to IPv4
AAAA: URL to IPv6
CNAME: URL to URL
ALIAS: URL to AWS Resource

Question 7

How do health checks relate to DNS Queries

Answer 7

Only healthy resources will be returned as a response to a DNS query. This enables failover routing. Health checks are mandatory for fail over R53 Policy

Question 8

can health checks be integrated with Cloudwatch?

Answer 8

Yes

Question 9

Why would you use a weighted routing policy?

Answer 9

To control the percentage of requests going to a specific endpoint. Helpful for:

• A/B Deployments
• Splitting traffic between regions

Question 10

What is the default health check interval in R53 and roughly how many health checkers will check and endpoint?

Answer 10

One every 30 second interval per health check with approximately 15 health checks per end point.

Question 11

By default how many health checks need to fail for an instance to be unhealthy? How many for it to become healthy?

Answer 11

3 for both

Question 12

When would you use a latency routing policy?

Answer 12

When latency for users is a priority. Under this policy traffic is redirected to the server that has the least latency as evaluated in terms of the user to the designated region.

Question 13

In a latency routing policy will traffic be directed to the closest region?

Answer 13

Not especially. The lowest latency will be used which may not be the closest region to the user

Question 14

Does R53 support healthchecks?

Answer 14

Yes - Health checks can be configured to simulate common user requests

Question 15

What are the 6 routing policies for R53

Answer 15

Simple, Multivalue, Geo-location, latency, failover, weighted.

Question 16

How are alias records resolved externally

Answer 16

As A or AAAA records

Question 17

In R53 where can private domain names be resolved

Answer 17

Private Domain names can only be resolved by instances within your VPC

Question 18

Does a simple routing policy allow for health checks?

Answer 18

No. You cannot attach a health check to a simple routing policy.

Question 19

When would you use a simple routing policy?

Answer 19

When you only need to redirect to a single resource

Question 20

If multiple values are returned from a simple routing policy, what action does the client take?

Answer 20

It picks a random value from the list

Question 21

Can a CNAME record point to a root domain - i.e. planittesting.com

Answer 21

No. A CNAME record can only point to subdomains - i.e. test.planittesting.com

Question 22

Why would you use a geolocation policy?

Answer 22

When you need to route traffic for specific users to specific regions based on their geographic location. You should always have a default policy if a region cannot be identified

Question 23

What is the purpose of a default policy in R53 geolocation and is it mandatory?

Answer 23

The default policy is used if a match cannot be found based on the users location. It is not mandatory.

Question 24

If you have a failover record-set in R53 called ‘failover’ which contains the IP for the primary and the secondary for the domain example.com - what is the URL for the primary and the secondary?

Answer 24

failover.example.com for both. R53 will switch from primary and secondary in the record-set automatically.

Question 25

If you need a heath check to check just the connectivity of a port - which sort of health check would you use?

Answer 25

TCP health check

Question 26

If you have an https health check does this verify the SSL certificate associated with the connection?

Answer 26

No

Question 27

You have a weighted routing policy setup. 50% of traffic goes to EU-WEST-1, 50% goes to US-EAST-1. You change the policy to 70/30 respectively and go to the browser but the old policy seems to still be in effect - whats the simplest reason?

Answer 27

Its likely that the TTL has not expired and your browser is still executing from cache

Question 28

Which layer does an application ELB work on?

Answer 28

Layer 7 - HTTP. Allows balancing on the URL. Allows balancing to multiple apps on the same instance.

Question 29

What protocols can be used in an R53 health check?

Answer 29

HTTP, HTTPS, TCP

Question 30

What are the 3 failover policies in R53?

Answer 30

Active/Active: R54 returns primary resource. In case of failure R53 returns backup resource. Must be a failover policy

Active/Passive: Multiple active resources. In case of failure return a healthy resource - any policy but failover

Combination of Both: Multiple routing policies built as a reee

Question 31

What the maximum number of endpoints returned in an MVA query?

Answer 31

8

Question 32

When would you use an MVA routing policy? Is this a replacement for an ELB?

Answer 32

When you want R53 to route traffic to multiple endpoints. It is not an ELB replacement