AWS EC2 Flashcards Preview

Learn AWS > AWS EC2 > Flashcards

Flashcards in AWS EC2 Deck (344)
Loading flashcards...
1
Q

List the EC2 instance categories?

A

Spot Instance, On-Demand Instances, Reserved Instances.

2
Q

What is an EC2 spot instance?

A

You can bid on unused EC2 capacity by using a spot instance, but a spot instance can be stopped and unallocated by AWS at any point in time.

3
Q

What is a reserved instance?

A

You pay upfront for EC2 capacity and you are guaranteed to be able to use this capacity when you need it, even if the AWS region is at 100% capacity.

4
Q

What is a on demand instance?

A

You use what you need and pay as you go.

5
Q

When using a reserved instance are you guaranteed you will be able to provision the EC2 instance when need, even if the AWS region is at 100% capacity.

A

Yes

6
Q

When using spot instances are you gaurenteed resources?

A

No

7
Q

When using spot instance can the instance be stoped at any time?

A

Yes.

8
Q

When using reserved instance can the instance be stoped at any time?

A

No.

9
Q

When using on-demand instance can the instance be stoped at any time?

A

No.

10
Q

With on demand instance are you gaurenteed resources

A

No.

11
Q

Is a spot instance the best choice for a situation where the load is changing all the time and the workload can not be interupted?

A

No a better choice here would be on-demand instance.

12
Q

What is AWS free tier?

A

When you sign up and during the first 12months you get some free resources with some restrictions, this is to help you learn.

13
Q

Is spot instances a lower cost then other instance types?

A

Yes as you bit for unused capacity.

14
Q

What sort of processes are best suited to spot instances?

A

Batch processes or tasks.

15
Q

If you know you will need a instance 24/7 365, what instance type is best suited to get best value, the work load can not be interrupted.

A

Reserved as it gives you the low price as you pay upfront and are guarantee the resources.

16
Q

`

A

No.

17
Q

When you buy a reserved instance is it in all availability zones?

A

No it is for a single availability zone.

18
Q

Can you transfer a reserved instances from one availability zone to another ?

A

Yes.

19
Q

If you have a large reserved instance and you do not need it any more but need a number of smaller instances, what can you do?

A

You can create a number of small instances from a large reserved instance.

20
Q

What is the default number of instances you can create ?

A

20

21
Q

Mow many elastic IP do you get by default?

A

5

22
Q

Where you reached the default max limits of EIP’s or Instances, what can you do to get more resources?

A

You can ask AWS to give you more resources, this typically takes 24hrs.

23
Q

Is ES2 instances fault tolerant?

A

No you have to architect a solution with fault tolerance.

24
Q

Dose ES2 instances integrate with IAM?

A

Yes

25
Q

What is the easiest way for application running on a EC2 instance to use the AWS service API’s with out needing IAM keys?

A

Create a IAM role for EC2 instance.

26
Q

What is EC2?

A

Elastic cloud compute.

27
Q

Are you paying for compute resource when the VM is stopped?

A

No.

28
Q

When a instances is stopped are you paying for any other resources used by the instance?

A

Storage used by EC2.

29
Q

What are the EC2 classic instances sizes?

A

micro, small, medium, large.

30
Q

If you what to stop some one stopping you instance by mistake, what can you do?

A

You can check the ‘terminate protection’ check box.

31
Q

What is a EC2 instance created from?

A

AMI’s.

32
Q

What are AMI’s?

A

An AMI is a sort of a template containing all the information required for creating one or more EC2 instances. The template(AMI) consists of,

  • Owner
  • Launch permissions
  • Architecture (64-bit X86 or 64bit-ARM)
  • Operating system
  • Block devise mapping
33
Q

Is it possible for you to create a custom AMI?

A

Yes, you can build a custom AMI for Linux and Windows.

34
Q

What services do you use to give EC2 instances volumes?

A

EBS.

35
Q

What do you pay for with EC2?

A

Compute time if the instance is powered on, storage, data transfer.

36
Q

There are currently two places a EC2 instance can store data, what are they?

A

Instance store and EBS.

37
Q

Where a instance is using a instance store and you stop the instance, will you data persist?

A

No when you stop a instance the data is gone/deleted.

38
Q

Where is the data stored when the EC2 instance is using a instance store.

A

Data is tired on the local had disks of the server and not remote storage.

39
Q

What is EBS?

A

Elastic block store.

40
Q

Where do EBS volumes live in relation to EC2 instance?

A

The volume is remote from the EC2 instance?

41
Q

What are provisioned IOPS in relation to EBS?

A

The ability to configure the amount of IOPS you required for you EBS volume.

42
Q

How would you calculate MB transfer per second?

A

IOPS + 16K chunks = MB transfer per second.

43
Q

What sort of application would benefit from provisioned IOPS?

A

Database applications, because of the random read and write patterns.

44
Q

What is the min for provisioned iops?

A

100

45
Q

What is the max provisioned IOPS?

A

64000

46
Q

What is the min size of the EBS volume?

A

1GB

47
Q

What is the max size of a EBS volume?

A

1TB

48
Q

What raid is used for a EBS volume?

A

Raid0

49
Q

Can you connect a EBS volume to more then one instance?

A

No.

50
Q

Is EBS a object store?

A

No it is a block store.

51
Q

Is a EBS volume local storage on the server where the EC2 instance is running?

A

No it is network attached storage and is remove from the EC2 instance.

52
Q

What is pre-warming a EBS volume?

A

This is where AWS will preform a erase on the volume the first time the volume is used and this has a large impact on performance.

53
Q

Is it possible to snapshot a EBS volume?

A

Yes

54
Q

Where are snapshots stored?

A

?????

55
Q

Do you pay for snapshot storage?

A

?????

56
Q

Are EBS snapshots incremental or full?

A

Incremental.

57
Q

Why would you take frequent snapshots?

A

To increase durability?

58
Q

Do snapshots degrade application performance what they are happening?

A

Yes.

59
Q

Will data persist on a EBS volume?

A

Yes.

60
Q

What Linux command would you use to touch every block on the volume during pre-warming?

A

dd command.

61
Q

How would you backup a EBS volume?

A

take a snapshot.

62
Q

How can you move a EBS volume to another availability zone?

A

??

63
Q

Can you schedule snapshots?

A

yes using the aws command line tools.

64
Q

What is user data used (script) used for, in reference to EC2.

A

This is a script that is run at the beginning of VM boot up.

65
Q

What is the cloud-init script used for, in reference to EC2.

A

This is a script that is run at the beginning of VM boot up.

66
Q

How can you from a script or applications access the EC2 features?

A

API or execute use the CLI.

67
Q

When you create a new EC2-Classic instance dose it get a public IP?

A

Yes

68
Q

When you create a new EC2-Classic instance dose it get a private IP?

A

Yes

69
Q

When you create a new EC2-Classic instance dose it get a CNAME?

A

Yes

70
Q

When you start or stop the instance will the private IP change, when using EC2-Classic?

A

Yes

71
Q

When you start or stop the instance will the public IP change, when using EC2-Classic?

A

Yes

72
Q

When you start or stop the instance will the CNAME change, when using EC2-Classic?

A

Yes

73
Q

How can you assign a static IP for a for a EC2-Classic?

A

Assign a elastic IP.

74
Q

Is a elastic IP a public or private IP?

A

Public.

75
Q

How many elastic IP’s can we assign to a EC2-Classice instance?

A

0ne per instance

76
Q

What is the default number of elastic IP you get?

A

You get 5.

77
Q

Can you request more elastic IP?

A

Yes.

78
Q

When using elastic IP’s and you shutdown a instance, will the elastic ip change?

A

No.

79
Q

Can you detach and reattach a elastic IP to a new EC2 instance?

A

Yes.

80
Q

When you stop a instance is the elastic ip detached

A

yes and you will have to reattach it to the instance.

81
Q

What is the elastic load balancer used for?

A

To distribute traffic between EC2 instances.

82
Q

Can the elastic load balance be used with other services?

A

No it is used for distributing traffic to EC2 instances in a ELB group.

83
Q

When using the ELB and a instance stops will ELB keep sending traffic to the failed instance?

A

No ELB will detect that the instance has failed and stop sending traffic.

84
Q

How do a ELB detact when an instance has faild?

A

ELB will send pings to the instance and after a treshold of faild pings the ELB will stop sending traffic to the faild instan

85
Q

Dose ELB store SSL certificates?

A

Yes ELB is used to store SSl certificates.

86
Q

Is it best practice to put sensitive data in the tags/metadata?

A

No as the tags/metadata is not encrypted.

87
Q

For EC2-Classic security groups is there filtering on outgoing traffic?

A

No.

88
Q

With regard to EC2-Classic, by default has the security group got port 22(ssh) open?

A

Yes.

89
Q

With regard to EC2-Classic, by default has the security group got ports other then 22(ssh) open?

A

No.

90
Q

Can you monitor a EC2 instance?

A

You can monitor the instance from a hardware perspective.

91
Q

What EC2 monitoring parameters do you get by default?

A

CPU, Disk Reads, Disk Writes, Disk read operations, Network in, Network out, status check failed instance, status check failed system, .

92
Q

Are the EC2 monitoring parameters the host parameters or the VM OS parameters?

A

Host parameters.

93
Q

With EC2 basic monitoring is the interval time 1min?

A

No the basic monitoring is 5min, this is the default monitoring.

94
Q

With regard to EC2 monitoring the default interval is 5min, can you get a faster interval limit?

A

Yes you can have 1min.

95
Q

With regard to EC2 monitoring the default interval is 5min and when you select a interval of 1min do you pay for it?

A

Yes you pay extra.

96
Q

Is it possible to create alarms with EC2 monitoring?

A

Yes you can create alarms and have them monitored by CloudWatch.

97
Q

Can you attach a volume in one AZ to a instance in another AZ?

A

No.

98
Q

Can you launch a volume from a snapshot?

A

Yes.

99
Q

How would you move a volume from one AZ to another?

A

Create a snapshot, create volume from image.

100
Q

How would you copy a volume from one AZ region to another AZ region?

A

Create snapshot, copy snapshot (this will gibe you region options)

101
Q

Do EBS volumes exist outside of a AZ and with in the region?

A

No, volumes exist with in the AZ (data centre)

102
Q

How can you create a image from a snapshot?

A

AWS provides a create image from snapshot function.

103
Q

How many 9’s availability has a EBS volume snapshot?

A

4 x 9s.

104
Q

Where are EBS volume snapshots stored?

A

On AWS S3 storage.

105
Q

How may 9’s durability has EBS snapshots?

A

11 X 9’s.

106
Q

Are snapshots full or incremental snapshots?

A

Snapshots are incremental.

107
Q

What is a provisioned IOPS EBS volume?

A

It is a volume where you have defined the IOPS for the volume.

108
Q

With provisioned IOPS EBS volume do you pay any other charges then the storage charge?

A

Yes you pay for the IOP’s.

109
Q

Can you attaches multiple volumes to a instance?

A

Yes.

110
Q

What durability has a EBS volume?

A

Data is replicated with in a AZ to ensure again a single disk failure.

111
Q

Dose amazon offer the ability to encrypt a EBS volume?

A

No.

112
Q

Dose Ubuntu support user data ?

A

Yes, user data is much like cloud init and is supported by Ubuntu.

113
Q

Dose amazon distro’s support user date?

A

Yes, user data is much like cloud init and is supported by amazon distro.

114
Q

What is user date?

A

It is a plane bash script that is run by supported distros during start up.

115
Q

What is a security group?

A

It is like a firewall.

116
Q

When security groups with EC2-Classic can you filter outgoing traffic?

A

No filtering is on input traffic only.

117
Q

Do instances with in the same security group require ports to be opened in the security group to enable inter instance communicate?

A

No instances in the same security group are free to send and receive traffic between other instances.

118
Q

Can instances in two different security groups communicate with each other by default?

A

No, instances in separate security groups require ports to be opened in the security groups.

119
Q

With regard to security groups, is it possible to allow all traffic one group to communicate with instances in another group?

A

Yes, you can use all t traffic and the group id instead of 0.0.0.0/0.

120
Q

What types of placement groups are they?

A
  • Clustered
  • Partitioned
  • Spread
121
Q

I have an HPC workload, how should I deploy this with a placement group?

A

Clustered, keeping the instances close together for high bandwidth, low latency. This way they get a 10G non-blocking network. But they are exposed to be on the same hardware and failure can take all instances down.

122
Q

What type of network is important when using clustered placement groups?

A

In clustered placement groups you get a 10GiB network thet is non-blocking and low latency.

123
Q

Are Clustered placement groups available across AZ’s?

A

No, all instances have to live inside a single AZ.

124
Q

If you are using provisioned IOPS, do you have to pay for them even if you do not use them?

A

Yes.

125
Q

What is the default interval that CloudWatch will monitor at?

A

5min.

126
Q

Is detailed CloudWatch monitoring enabled on a per instances basis?

A

True.

127
Q

Is it AWS best practice to store secret/access keys on the AMI?

A

No.

128
Q

With regard to EC2-Classic, can you change a instances security group with our removing the instance?

A

No.

129
Q

With regard to EC2-Classic, how can you change the security group?

A

You can not, but you can boot a instance from existing AMI into the new security group.

130
Q

Can a instance be a member of more then one security group?

A

Yes.

131
Q

With regard to EC2-Instance, will the EIP become detached when the instance is stopped?

A

Yes

132
Q

With regard to EC2-Instance, when you stop and start a instance will the EIP automatically get reattached?

A

No you have to manually reattach it.

133
Q

Is EC2 a global, region or availability zone service?

A

It is a availability zone service.

134
Q

When a instance is in a VPC and it is stopped and started dose it loose it’s Elastic IP?

A

No.

135
Q

When you start and stop and instance in a VPC, will the instance lose it private iP?

A

?????

136
Q

What is Intel’s AES-NI?

A

AES-NI is 7 new Intel instructions for AES.

137
Q

What is Intel’s DRNG?

A

It is a digital random number generator in the CPU, used with encryption.

138
Q

What are placement groups?

A

Placement groups control how EC2 instances are place, you can have 3 different types,

  • Clustered
  • Spread
  • Partationed
139
Q

When selecting a reserved instance, what are the attributes that apply

A
  • Instance type
  • Scope: Wheather applies to a region or a AZ
  • Tenancy: shared or single
  • Platform: Windows or Linux
140
Q

I have a reserved instance they expired, will the instance stop running?

A

No, the instance will keep running but is billed as on-demand rate.

141
Q

Is a reserved instance a physical or virtual instance reserved for your usage?

A

No, it is as billing construct applied to your account.

142
Q

What is a standard and convertible instance?

A

It is used when reserviceng an instance, a standard instance can miodify AZ,Scope, network,platfrom and instance size with ins same instance type. A convertable iinstance can

143
Q

Can I sell a standard reserved instance in the market place?

A

Yes

144
Q

Can I sell a convertible reserved instance in the market place?

A

No

145
Q

What is a reserved instance?

A

You make a 1 or 3-year commitment and receive a significant discount over on-demand. Up to 75%.

146
Q

What are reserved instances good for?

A

Stady state workloads.

147
Q

What is a reserved instance with a capacity reservation?

A

This is where you reserve the capacity, you are guaranteed to be able to create an instance when needed.

148
Q

What are the upfront option available?

A
  • No Upfront
  • Partial upfront: Up to 76%, you pay only for the hours after the up front
  • All upfront, you do not pay for any hours during term
149
Q

I have five EC2 instances that will share huge amounts of data between them, how should I place them?

A

You will need a low latency network thet is also high speed. Clustered placement groups give you 10GB bandwidth and low latency. All in a single AZ.

150
Q

What is placement group - cluster?

A

Cluster – creates a cluster of instances in a single AV with a 10GiB network and low latency non-blocking.

151
Q

What is placement group - partition?

A

Spreads instances in groups across racks. Multiple AZ.

152
Q

What is placement group - spread?

A

A spread placement group is a group of instances that are each placed on distinct racks, with each rack having its own network and power source. Multiple AZ.

153
Q

When using a clustered placement group, what are the attributes I need to think of?

A

Single-AZ
10Gpbs non-blocking
Means Instances will be place in same rack.
Instance type need to have enhanced networking to make bets use

154
Q

I have a workload that requires the three hosts to have the hastes, lowest latency network possible, we are not worried about availability or durability of the workload, what is my bets poption?

A

Use clustered placement group, the cluster placement group will cluster the hosts logically so they have 10GB non-blocking network.

155
Q

I need to spread my instances across a single AZ in groups, but I want to ensure the instances in the groups are not on the same hardware, what is my best option?

A

Use partition placement groups, this will enable you to spread across a single AZ and no instance will be placed on the same host and another instance.

156
Q

When using placement groups, what do I need to be aware of?

A

The hardware available will be of limited size, you may when launching gets an error, just wait and try again.

157
Q

I want to ensure my instances do not share the same hardware dor availability purpus, I wnat to place the instances in two AZ, what options do I have

A

Spread placement groups, spread placement groups enable you to spread instances and no two of the cases will be on the same hardware, also these can be spread across two AZ.

158
Q

What is the limit size of spread placement group?

A

7 Instances.

159
Q

Can I have placement groups across regions?

A

No

160
Q

Can I have placement groups across AZ

A

Yes but only spread placement group.

161
Q

I want to use a T1 micro with a placement group is there any issues?

A

Yes, clustered PG are high-performance networking and require an instance with enhanced networking, but spread PG will work.

162
Q

I need to create a copy of an instance in another region, can I just snapshot and copy?

A

Yes, create an AMI and copy the AMI to another region.

163
Q

I need to copy an AMI to another region, is this possible?

A

Yes

164
Q

I have an instance running as on-demand and default-tenancy, I want to change this instance to single-tenant, how cna I do this?

A

You have to stop the instance to ensure no changes are happening to the instance, create an AMI and launch a new instance.

165
Q

What is the reserved billing model?

A

You can reserve an instance like RDS, EC2, DynamoDB.

166
Q

I need 100% guarantee that I will be able to have capacity available when I need to start up and instance, what options do I have?

A

You cna use availability zone reserved instance, this will guarantee that I can spin up instances in AZ when needed.

167
Q

What is the difference between availability and regional reserved instances?

A

Availability reserved instance give you a guarantee that you can have the capacity to sping up an instance, regional reserved does not.

168
Q

What are the payment models for reserved instances?

A

All upfront, partial, and no upfront.

169
Q

What is the 3 billing models for instances?

A
  • On demand
  • Reserved
  • Spot
170
Q

For reserved billing model, what is the available billing terms?

A

12 or 36 months

171
Q

I am using spot instance and I need to get notified when an instance is been reclaimed by AWS, what are my options?

A

You get a termination notice from the metadata server here, http://169.254.169.254/latest/meta-data/spot/termination-time

172
Q

Can I have multiple network interfaces?

A

Yes

173
Q

Are security groups associated with the instance or the ENI (network interface)?

A

ENI

174
Q

What is the first ENI on an instance called?

A

Elastic Network Interface.

175
Q

What is the root volume of an EC2 instance?

A

It is the volume used for booting off of.

176
Q

I wnat to create an AMI for my instance, can this be done directly in the API or portal and if so what sequence of events happen in AWS?

A

Yes, you can create an AMI direct in the portal using Create Image. This will create snapshots of the instance root volume and any attached volumes. The AMI will container the information for the size, volumes, etc. The AMI will reference the snapshots created, the AMI does not contain the volume/snapshot data.

177
Q

Is the AMI a regional or global resource?

A

AMI’s are regional resources if you wnat a copy of the AMI in another region you can copy it to the other region.

178
Q

When you copy AMI to another region, what happens in AWS?

A

AWS will copy the AMI (template not the data ) and it will copy each of the volumes (snapshots). When finished you will have, the AMI and a snapshot for each of the volumes.

179
Q

What are AMI permissions?

A

You can make the AMI public or private or share with an anoyther account,

180
Q

I wnat to share my AMI with another account, can I do this?

A

Yes using AMI permissions.

181
Q

What am I doing when I set an AMI permission to the public?

A

You are sharing your AMI with the AWS AMI community.

182
Q

I have a really cool todo list app that I have packaged as an AMI, how could I sell this AMI?

A

You can use the AMI market place, you add a price so the total price to the user is the AWS resource cost and your cost for the software.

183
Q

I need bare-metal performance for an instance, what options do I have?

A

Two options,
- Bare metal
-Nitro
Nitro is probably the better choice.

184
Q

I need to get better network and storage performance what choosing instance technology, what options do I| have?

A

SIROV

185
Q

What types of storage are available to an instance?

A
  • Instance store
  • GP2
  • IO1
  • St1
  • Sc1
  • Provisioned IOPS
186
Q

I need to store tempory data fast for an instance, what is my best option?

A

Instance stored volumes where I can store the data on the disk (SSD) on the instance the hardware is running on.

187
Q

What type of storage is instance storage?

A
  • Tempory
  • On the same hardware as an instance
  • NVMe
188
Q

I need real fast tempory storage, it needs to be higher throughput than a single instance store, what options do I have?

A

RAID a number of instances stores

189
Q

I need to ensure that my EBS storage has consistency, what options do I have?

A

Use storage optimised instances as they have dedicated storage path on the network

190
Q

I need shared storage, I am thinking of EBS, is this the correct storage?

A

No, EBS is attached to a single instance only and can not be shared, EFS could be an option.

191
Q

I am thinking of using EBS for high durability requiring 99.9999% storage, is EBS a good choice?

A

No, EBS has only 99.95 - 99.99% durability.

192
Q

Is a snapshot full or incremental?

A

Incremental.

193
Q

I need to ensure that my instance is recoverable in the event of a failure, what options do I have?

A

Take a snapshot, the snapshot is stored in s3 and as such has a higher degree of durability then EBS.

194
Q

I needed to increase the durability of my EBS volume, how can I do this?

A

Take a snapshot (backup) and store it in s3, in S3 data dat is across AZ’s and 11 9s durables.

195
Q

Is snapshot consistent?

A

No, you have to do one of two things, you have to shut the instance down or unmount the volume.

196
Q

I wnat to move my instance between AZs, how can do this?

A

You will have to take a snapshot of the instance EBS volume and then use the snapshot to create a new instance in the other AZ.

197
Q

I am thinking of using EBS for static content, is this a good option or what else should I use?

A

Do not use EBS, use S3.

198
Q

How do the IAM role credentials (accesskey, securitykey) get to the application on the instance?

A

When you assign a role to an instance, the STS creates a set of creds (access + security keys), these are pushed to the metadata server, your app can get these form the metadata server. To get the STS, the instance service is assuming the role and in so getting the STS creds, what is happening is there is a trust relationship for the role between the role and the ec2 instance.

199
Q

When would I use an IAM Role with an EC2 instance?

A

Anytime your application running on the EC2 instance wants to call other AWS services API.

200
Q

What types of placement groups can we have?

A
  • Cluster PG: In a single AZ, Highspeed low latency
  • Petition PG
  • Spread PG
201
Q

What characteristics has a Clustered Placement Group?

A
  • For the highest network performance as all VM’s are close to each other as possible.
  • Latency is a minium
  • Placement group is in a single AZ
  • Susceptible to failure of AW or hardware as this could be on a single rack, server and is in a single AZ
202
Q

I have a workload that requires the ability to calculated a genome in a distributed fashion, how can I make this happen using EC2 instance and what will the configuration be?

A
  • Large compute optimised instance type like c5n.18xlarge
  • Clustered placement group to ensure network latency is a minimum
  • Ensure the workload can deal with failure as the instances will be in close proximity to each other.
203
Q

I am creating a clustered placement group, when provisioning this clustered PG, what possible error cloud I get from AWS?

A

You could get an error in relation to these not being enough resource for your request. This is due to the close proximity of resource needed in a placement group.

204
Q

I wnat to add instances to an existing clustered placement group, how cna I do this?

A

You cna add instances to an existing PG but you may get capacity issues as AWS provisioned everything up front. You are best to delete the CPG and recreate it.

205
Q

Should you mix instances type in a clustered placement group?

A

No.

206
Q

I need to pick a placement group for performance, what type of placement group should I pick?

A

Clustered PG as it keeps the instances close to each other physically giving best network performance and latency.

207
Q

I need to puck a placement group for resiliency, what type of PG should I use?

A

You should use spread PG as all the instances will be spread across AZ’s and ensure they are as far apart within the available AZ’s

208
Q

What is a spread placement group?

A

It is a PG that spreads you instances across AZ’s and ensure they now VM’s are on the same hardware if possible. Does this give the highest level of resiliency?

209
Q

Which of the placement groups gives you the best resilience and why?

A

Spread placement groups ensure your instances are spread across AZs and kept on separate hardware.

210
Q

What is a partition placement group?

A

This will spread SETS of instance across AZ’s. The SETS are grouped close but each set is distributed across the AZ’s.

211
Q

I an add instance sot spread placement groups after the SPG is deployed?

A

Yes

212
Q

I need to have 21 instances use a spread placement group, how should this be configured?

A

You will have to have 3 AZs as an SPG can only have 7 instances per AZ.

213
Q

Can you use a VPC peer with Cluster Placement Groups?

A

Yes

214
Q

Can I deploy clustered PG across AZ’s?

A

No.

215
Q

I need 10GB flow for traffic, what is my best option for deploying 3 instances for performing a calculation?

A

Clustered PG as you cna get 10GB traffic flow.

216
Q

What type of networking does a cluster placement group get?

A

As all instances are physically close in the same AZ, you get 10GB, no-blocking, non-oversubscribed.

217
Q

How does AWS deploy partition placement groups?

A

AWS groups you instance sin to partition, so where you have 9 instance you get 3 partition of 3 instances. Each of the 3 instance partitions is put on separate rach with own pwr, network, etc. Partitions are spread across AZ’s within the regions. No two partitions share the same rack.

218
Q

What is the Data Life Cycle Manager

A

DLM enables you to create a policy that enables backup of your EBS volumes.

219
Q

For Data Life Cycle Manager how do you create a policy and apply it to a volume?

A

You create a policy, define one or more tags, a schedule,

220
Q

I wnat to backup by EBS volumes, in my enterprise, do I create a CRON job or something else?

A

You can use the Data Life Cycle Manager to backup your EBS volumes, you define one or more volume tags and schedule.

221
Q

I have an online t-shirt web store, the number of customers is extremely variable and near impossible to predict workload, I have to select a billing model for my EC2 instances, what are my best options?

A

You should select on-demand as you do not know the workload from hour to an hour and day to day.

222
Q

I have an online t-shirt web store, the number of customers is steady at a base lone and also peeks, how should I select a billing model?

A

Select reserved for the baseline and on-demand for the peeks. If the web apps can deal with been terminated use spot instead of on-demand.

223
Q

I need to be able to ensure that I can always start my EC2 instances even if an AZ fails or there is high demand for instances in AWS, what is my best option?

A

Reserved will ensure you can always start your instances.

224
Q

What is a scheduled reserved instance?

A

This is where you receive the instance by a day or hour if you are doing it regularly.

225
Q

What is a spot instance pool?

A

Is a group of instances of the same type, a spot pool is used with a spot fleet

226
Q

How much can I save on the cost of a spot instance?

A

60 - 90%

227
Q

How much can you save on reserved instance?

A

60% (3year commit)

228
Q

How does spot pricing work/

A

After MAR 2018 you just pay the spot price in effect for each hour.

229
Q

What is EC2 fleet and how does it work?

A

EC2 Fleet enables you to define a capacity that you would like, this can be in the form of any instance sizes and use spot or on-demand. EC2 Fleet will attempt to create a pool of instance according to your desired capacity.

230
Q

I wnat to have a pool of computing as a baseline for number crunching, some times we need results faster and we would like to be able to take advantage of lower pricing spot instances, what can I do?

A

You cna use fleet to define the pool fo ec2 instance you wnat and when needed change the fleet config to use spot instances when the spot price is where you are willing to pay for it.

231
Q

What is AMI baking?

A

This is where we create an AMI with everything you need baked-in, like a web server and its configuration-

232
Q

If I have an autoscaling group and I wnat to be able to very quickly spin up instances, I can wait for use data scripts to execute as the lag would defeat the requirement to sping up quick, what are my options?

A

You can use AMI baking, this is where you create an AMI with the apps such as web servicer and the configuration into the AMI.

233
Q

What are the two components of an autoscaling group?

A

Launch template and autoscaling group

234
Q

Can I use spot instances with Autoscaling groups?

A

Yes 100%, you define it in the launch config.

235
Q

I have an already deployed Autoscaling group with launch config, I needed to make changes to the launch config, can I just edit the launch config?

A

No, you have to create a new launch config.

236
Q

Can an autoscaling group have a second launch config?

A

No

237
Q

Whts is the difference between the launch config and launch template?

A

Launch templates are the newer way, you have lots more

238
Q

What happens if an instance in an autoscaling group fails?

A

The instance is replaced.

239
Q

What is health check grace?

A

It is the time waited before an autoscale group checks the health of an instance, it gives instance time to build and become stable.

240
Q

I wnat to know when an autoscaling group changes, how cna I achieve this?

A

SNS can use with Autoscaling groups.

241
Q

Can I manually edit the auto scaling group to do things like the increase desired number of instances?

A

Yes.

242
Q

What is a scaling policy?

A
Scaling policy enables you to use metrics to scale up or down.
Metric types,
- CPU Utilization
 -Net in or Out
- Load balancer request count
243
Q

What are the AutoScaling policy types?

A

Simple
Scaling steps
Tracking

244
Q

Explain simple AustScaling policy type?

A

The simple policy type enables you to react to

245
Q

Explain the target tracking AustScaling policy type?

A

With target tracking, you set what you wnat the metric (CPU) to be an AS will add or remove instances to try to match it.

246
Q

Between 2 - 3 pm each day my web server workload increase, what would be a good option for autoscaling so as not to have a delay when loads start to increase?

A

Create a schedule and have the autoscaling group increase it number f instance in advance and then scale back down after the event has finished.

247
Q

I am building a set of web servers and autoscaling and I wnat to be able to take advantage of both spot and on-demand instances, is this possible?

A

Yes, you can when creating autoscaling opt to use the fleet model of using both spot and on-demand instances.

248
Q

I am using an auto scaling group, I require a notification to be sent what action like scale up occurs, how can we architect this?

A

Autoscaling has the ability to set up SNS to send a notification for,

  • Launch
  • Terminate
  • Fail to launch
  • Fail to terminate
249
Q

I wnat to get a notification when the autoscaling group fails to launch so we can have a ticket raided in our ITSM, how can I configure AutoScaling for this?

A

AutoScaling has a notification where we can configure notifications for the failed launch of instances to be sent to SNS.

250
Q

I wnat to get a notification when the autoscaling group fails to terminate so we can have a ticket raided in our ITSM, how can I configure AutoScaling for this?

A

AutoScaling has a notification where we can configure notifications for failed termination of instances to be sent to SNS.

251
Q

Explain how we can monitor AutoScaling groups?

A

We cna enable group metrics for the autoscaling group

252
Q

I wnat an Autoscaling policy to react based on how much the alarm is our by what would I need?

A

Auto-scaling with steps.

253
Q

If I have 6 subnets and I deploy 12 instances, how are the instances distributed when using an autoscaling group?

A

All instances are distributed evenly, each subnet will get 2 instances.

254
Q

I am using 3 AZs and an autoscaling group, I require a total of 12 instances to be running to service my workload, a failure of an AZ occurs, the Autoscaling group attempts to spin up the lost four instances in the remaining AZ’s, but it gets our of capacity error, what can I do to fix this in the future?

A

Purchase reserved instances, another option is to add most AZ’s and this way you lose less of your instances when an AZ goes down, you also get a better chance of finding space in the remaining AZ’s

255
Q

What is AWS Data Lifecycle Manager?

A

It allows you to schedule EBS snapshots

256
Q

With AWS Data Lifecycle Manager, can I set a retention policy?

A

Yes

257
Q

When using AWS Data Lifecycle Manager, what is the mandatory action I have to do with the EBS to get AWS Data Lifecycle Manager to work?

A

You have to add a tag so AWS Data Lifecycle Manager knows what to match the volumes it will snapshot.

258
Q

I need to take EBS snapshots on a regular basis, once every 12 hrs, I am writing a script for a cron job, is this the bets solution?

A

No, it is better to use AWS Data Lifecycle Manager

259
Q

For reserved instances with is the therm?

A

12 and 36mts

260
Q

For reserved instances what is the upfront models available?

A

-All Upfront
Partial
No upfront

261
Q

I require an instance but from day to day and week to week I do not know if I will need it, one other factor is when it is running I can not have it stoped for any reason, what are the bets instance billing model to use?

A

Use on-demand as you can use what you wnat when you need it and stop using instance when you do not need it.

262
Q

What is the priority given to the EC2 instance billing model when starting n up?

A
  • Spot = low priority
  • On-demand = natural priority
  • Reserved capacity= high priority
263
Q

Is the reserved billing model only for EC2 instances?

A

No, it can be used for DynamoDB and also RDS instances.

264
Q

When would I use reserved?

A

When I can predict my usage.

265
Q

When I am acquiring a reservation, what is it tied to an AZ or a region?

A

Do you have the option of both?

266
Q

I need to ensure thet when I need an instance it will always be able to start, what do I need to do?

A

You need to do a reservation and use an AZ, using a region will not guarantee capacity and be able to start up.

267
Q

What types of hypervisor/physical options do I have with EC2?

A
  • ParaVitrtual
  • HVM (Hardware Virtual Machine)
  • Nitro
  • Baremetal
268
Q

What are the EC2 families?

A
  • i : Storage optimised
  • T + M @ General purpose
  • C : Compute-optimized
  • F : FPGA
  • G : GPU
  • R : Memory optimised
  • A : ARM instances
    = P Accelerated computing (8 NVIDIA Tesla cards, CODO)
269
Q

When would I use a C instance type?

A

When you need a compute-optimized workload.

270
Q

When would I use I instance type?

A

When you need a storage optimized workloads.

271
Q

When would I use an R instance type?

A

When you need a memory-optimized workload.

272
Q

When would I use an M instance type?

A

For general workloads with consistent CPU usage

273
Q

When would I use a T instance type?

A

For a general workload with bust cup usage.

274
Q

When would I use a G instance type?

A

When I need a general workload with bursty CPU.

275
Q

I have a general-purpose workload what instance type should I pick?

A

M or T, M got consistent CPU and T for bursty.

276
Q

I have a bursty CPU workload, what type of instance should I pick?

A

T instance type.

277
Q

In m5a,large, what is the ‘a’ telling us?

A

It is telling us that this is using AMD and not intel.

278
Q

What way so the T instances work?

A

You gain credits when below the CPU threshold and you loose credit when above. When all credits are gone you cna have one of two things happen, performance is affected, or you pay a surcharge.

279
Q

I am using T instance type, what are the two options when you are out of credits?

A
  • You pay a surcharge

- Performance degrades

280
Q

Explain how T type instances work?

A

The accumulate credit when below the threshold set for the instance type for CPU and use credit when above the threshold.

281
Q

What are P instance types used for?

A

CUDA/NVIDIA

282
Q

What is CRIMP-TGF?

A
C Compute optimised
R Memory Optimised
I storage optimised
M General purpose
P Accelerate for CUDA/NVIDAI

T is good for burst and is a general instance type, you have two options, when you are out of credits you can have performance degraded or you cna have a surcharge
G UPU
F FPGA

283
Q

Do all instance type have instance storage?

A

No

284
Q

What type of instance storage can you have?

A

NVME and SSD

285
Q

What are instance storage type good for and why?

A

Instance storage is good for cache volume and transient storage, it is the highest speed read and writes..

286
Q

How many instance volumes can I have?

A

0 - 24

287
Q

If I do not have enough storage in a single instance volume how cna I add more?

A

You can RAID the instance storage volumes together to make a single combined volume. RAID 0 is best, no parity or mirror…

288
Q

If I start and stop an instance will I lose the instance volume data?

A

Yes

289
Q

If I restart the instance will I lose the instance volume data?

A

No

290
Q

How cna I have the instance move to a new host?

A

Restart the instance.

291
Q

I wnat to use shared storage, should I use instance storage?

A

No, because instance storage is transient.

292
Q

I wnat to use persistent storage, should I use instance storage?

A

No, because instance storage is transient.

293
Q

I wnat to use elastic storage, should I use instance storage?

A

No, because instance storage is transient.

294
Q

I wnat to use durable storage, should I use instance storage?

A

No, because instance storage is transient.

295
Q

What is an EC2 instance profile and where is it used?

A

An instance profile is a container for an IAM role that you can use to pass role information to an EC2 instance when the instance starts.

296
Q

How many roles can I associate with an EC2 instance?

A

1

297
Q

What types of placement groups are available?

A
  • Clustered
  • Partitioned
  • Spread
298
Q

I have an application thet runs on multiple instances and requires very low latency between instances, what is my best option?

A

Opt to use ‘clustered placement group.’ This will keep the in the same physical; location.

299
Q

When would we use spread placement group?

A

Designed for a small number of instances across a single-az or multi-AZ, will spread out the instances evenly across the AZ.

300
Q

How may AZ is a clustered placement group spread across?

A

A clustered placement group is located in just one AZ.

301
Q

I have a placement group, and I need to add a node, how do I do it?

A

You can modify a placement group, you have to delete and replace.

302
Q

I need extreme performance, as in my apps dose computation and the result between each node is share, requiring huge bandwidth between node, what is the best placement group type?

A

Clustered

303
Q

I require resilience in my nodes and what to ensure they are always spread across AZ, what placement group type should I use?

A

Spread

304
Q

When creating placement groups can I add instances ad-hoc when I need them?

A

No, you have to allocation all instance at the same time

305
Q

I am using a placement group am I guaranteed to get all instances in the group at the time of provisioning?

A

No, as the instance for clustered placement group are going to be close proximity to each other, you may not have enough resources available at d provisioning time.

306
Q

I need to add 5 new instances to an existing placement group, how can I do this?

A

You can not modify a placement group.

307
Q

Do all instances support the placement groups?

A

No

308
Q

I need to select a placement group for performance, what type of placement group should I pick?

A

Clustered

309
Q

What is a partition placement group?

A

Where AWS creates a number of partitions and groups the instances in the partitions equally.

310
Q

What is a spread placement group?

A

This is for a small number of instances spread across the largest number of different physical hardware.

311
Q

Is a clustered placement group restricted to a single VPC?

A

No, it can space VPC’s, but not availability zones, this would mean the VPC all have to be in the same AZ.

312
Q

What is the main benefit of a clustered VPC?

A

10Gbit non-blocking, non-oversubscribed, bi-directional networking.

313
Q

I have to create an HDFS, HBase, and Cassandra what placement group would use?

A

You would use a partitioned placement group.

314
Q

Explain how AWS creates partitioned placement groups?

A

They create a number of partitions and place racks in the partitions across AZ’s in a single region they then evenly distribute the instances.

315
Q

What is the max number of partitions in the partitioned placement group?

A

7

316
Q

I need to have the max availability and at the same time what to launch different instances at different points in time, what placement group would I use?

A

You would use a spread placement group, this will spread

317
Q

When using EC2 is it a regional, global or AZ’s product?

A

AZ, it is an AZ product and if an AZ was to fail you would lose your instance.

318
Q

What does an AMI consist of?

A
  • One or more EBS snapshots, or, for instance-store-backed AMIs, a template for the root volume of the instance (for example, an operating system, an application server, and applications).
  • Launch permissions that control which AWS accounts can use the AMI to launch instances.
  • A block device mapping that specifies the volumes to attach to the instance when it’s launched.
319
Q

Can I copy an AMI to another region?

A

Yes

320
Q

When I create an AMI, do I get all the block devices and their mappings?

A

Yes, a snapshot is taken of each EBS black device and the AMI also keep the block device mapping.

321
Q

Can you have Clustered placement groups across AZ?

A

No

322
Q

Can you have Partitioned placement groups across AZ?

A

Yes

323
Q

Can you have Spread placement groups across AZ?

A

Yes

324
Q

Can you have Clustered placement groups across peered VPC?

A

Yes

325
Q

How much discount fo you get from spot instances?

A

Up to 90%

326
Q

How much discount do you get form on-demand?

A

Zero

327
Q

How much discount do you get from RI?

A

40% - 60%

328
Q

How much discount do you get from spot?

A

50% - 90%

329
Q

How can I create a 1mi CPU instances?

A

You will need to use fleet, the fleet will enable you to target different capactty pools in a single region, this way you will be able to get the max number of instances.

330
Q

How can you maintain a set of spot instances?

A

You have this option, you can select ‘request and maintain ‘ in the console

331
Q

Explain spot instances?

A

You ask for a pool fo spot instances and this is given to you, you get the current spot price, this price can be up to 70% discount.

332
Q

I have a soilution and I wnat to ensure my instance recovers and keeps the same ip, how cna I do this?

A

When an instance fails and is restored, it will keep the same private IP. to recover we have to think about autoscaling form cloud-native or use cloud watch and StatusCheckFailed_System metric and use it to recover the instance.

333
Q

Will the IP of an AWS recovered instance change?

A

No, the same instance IP, instance name and other parameters all say the same including the id.

334
Q

How can I detect when an EC2 instances fails at the AWS level?

A

StatusCheckFailed_System will get triggered.

335
Q

What is affinity host?

A

Affinity allows you to specify which Dedicated Host an instance will run on after it has been stopped and restarted. This gives you the confidence that the instance will run on the same physical server even through planned interruptions, helps reduce licensing costs in scenarios that require license affinity for a period of time (e.g., 90 days), and can be used to maintain an instance placement scheme.

336
Q

Due to licencing, I wnat to control the placement of my software in the server, how cna I do this?

A

You can use a dedicated EC2 instance, you can define where (server)the software is placed.

337
Q

What is EC2 dedicated?

A

This is where you get a whole server dedicated to you, you still are running EC2.

338
Q

What are the advantages of EC2 dedicated?

A

You get to have a server thet you and you only run on, you can have it so when you stop and start instance it will always be on the same server.

339
Q

Is Ec2 dedicated bare metal?

A

No, the instances that run on a Dedicated Host are the same virtualized instances that you’d get with traditional EC2 instances that use the Nitro or Xen hypervisor.

340
Q

How does an EC2 dedicate how get allocated?

A

You allocated the host to the account, you can then create EC2 instances and have them deployed to the dedicated host. Or you can use the AWS Licence manager to have the software run on the dedicated host.

341
Q

Can you use reservation with EC2 dedicated?

A

Yes

342
Q

Can you use a service plan with EC2 dedicated?

A

Yes

343
Q

Can I mix EC2 instance types on a EC2 dedicated server?

A

Yes

344
Q

I need to ensure my MSSQL licence is always run on the same server, I can use RDS, what options do i have?

A

You cna used EC2 dedicated and you get a server thet is dedicated to you EC2 instances. You cna define the MSSQL server is only run on this server and this can be done with AWS Licence manager.