Day 3 · S4 EC2 Instance Storage · S5 ELB & ASG · S6 Amazon S3 There are 83 dump questions today: 19 for S4, 24 for S5 and 40 for S6. Each answer sits right under its question, so cover it and commit to a letter before you read it. The dump's key is wrong twice today, on Q290 and Q606. Most of this material is Domain 3 (34%). The shared-responsibility tables count toward Domain 2, and elasticity and HA count toward Domain 1. S4 · EC2 Instance Storage Storage options Service Type Attaches to Data survives stop/terminate? Exam keywords --------------- EBS volume Block storage on a network drive, locked to one AZ One instance at a time (multi-attach isn't tested) Yes. The root volume is deleted on termination by default "persistent block storage", "boot volume" EC2 Instance Store Block storage on a physical disk in the host Only its own instance No. It's ephemeral, and data is also lost if the hardware fails "highest I/O", "buffer/cache/scratch", "temporary" EFS File storage (managed NFS), multi-AZ Hundreds of Linux instances Yes "shared file system", "NFS", "Linux", "mount on many instances" FSx for Windows File Server File storage (SMB, NTFS), integrated with Microsoft AD Windows instances and on-prem clients Yes "SMB", "Windows", "Active Directory", "mapped drive" FSx for Lustre File storage for HPC (Lustre = Linux + cluster) Linux compute clusters; can link to S3 Yes "HPC", "ML/analytics", "100s of GB/s, millions of IOPS" FSx is AWS's family of fully managed third-party file systems. NetApp ONTAP is also part of it. EBS rules An EBS volume is locked to its AZ. A volume in us-east-1a can't be attached to an instance in us-east-1b. To move it, take a snapshot and restore the snapshot in the other AZ. You're billed for provisioned capacity (GB and IOPS) whether you use it or not. You can grow a volume later. Delete on Termination is on by default for the root volume and off for extra volumes. Turn it off on the root volume if you want to keep it after termination. Snapshots, AMIs and EC2 Image Builder Feature What it does Facts to know --------- EBS snapshot Point-in-time backup of a volume Detaching first is recommended but not required. Can be copied across AZs and Regions Snapshot Archive Cheaper tier for snapshots you rarely restore 75% cheaper; a restore takes 24–72 h Recycle Bin Recovers accidentally deleted snapshots Retention rule of 1 day to 1 year AMI A customized image of an instance (OS, software, config, monitoring) Faster boot. Tied to a Region but can be copied to others. Sources: public (from AWS), your own, AWS Marketplace EC2 Image Builder Automates building, testing and distributing AMIs (and container images) Runs on a schedule, distributes to multiple Regions, and is free apart from the resources it uses AMI process: launch an instance and customize it → stop it so the data is consistent → create the AMI, which also creates EBS snapshots → launch new instances from the AMI. EFS details EFS costs about 3× EBS gp2. You pay per GB used, with no capacity planning. EFS-IA is a storage class for files that aren't read every day, and it's up to 92% cheaper than EFS Standard. A lifecycle policy moves files there after N days without access (60, for example), and applications don't notice the move. One Zone classes keep data in a single AZ, which is cheaper but less resilient (see Q147). Shared responsibility for EC2 storage AWS You ------ The infrastructure Backup and snapshot procedures Replicating data for EBS volumes and EFS drives Data encryption Replacing faulty hardware All data on the drives Making sure AWS employees can't access your data Understanding the risk of using instance store Exam traps Storage types: EFS and FSx are file storage, EBS and instance store are block storage, and S3 is object storage. Storage shared by many instances is EFS for Linux/NFS or FSx for Windows for SMB. It's never EBS. "Ephemeral", "lost when stopped", "highest IOPS" or "scratch" means instance store. "Instance template prepackaged with software" means an AMI. "Automate building and testing AMIs" means EC2 Image Builder. Amazon Data Lifecycle Manager, which schedules snapshots, appears as a distractor. Disaster recovery for EC2 means AMIs plus EBS snapshots, copied to another Region. Who encrypts or backs up EBS: the customer. "Direct access over the internet" means S3, not EFS (Q290). Coming from on-prem On-prem AWS ------ VMDK on a datastore EBS volume Local SSD/NVMe in the ESXi host Instance store Veeam restore point (a real backup, unlike a vSphere snapshot) EBS snapshot Golden VM template AMI Monthly template refresh (patch, test, publish) EC2 Image Builder Linux NFS export EFS Windows file server with SMB shares and AD FSx for Windows File Server S4 questions (19) I've corrected OCR typos from the source and flagged other glitches. Q20. Amazon Elastic File System (Amazon EFS) and Amazon FSx offer which type of storage? A. File storage B. Object storage C. Block storage D. Instance store Answer: dump…