EFS, FSx and shared storage: SOA-C03 study notes and practice questions
What to know about EFS, FSx and shared storage for the SOA-C03 exam, from the Monitoring, Logging, Analysis, Remediation, and Performance Optimization domain (22% of the exam guide), then 1 free question to test it. Part of the SOA-C03 study guide.
Shared storage questions ask which file system fits, how to tune it, and why a mount fails.
Amazon EFS
NFS file system for Linux instances, containers and Lambda, growing and shrinking automatically.
Regional (multi-AZ) or One Zone file systems; use a mount target in each AZ you use.
Throughput modes:Elastic (scales automatically, pay for what you use; the default and usual answer), Provisioned (set throughput independent of size), Bursting (scales with size, uses credits).
Performance modes: General Purpose (lowest latency, the default) and Max I/O (older option for highly parallel workloads).
Storage classes: Standard, Infrequent Access and Archive; lifecycle management moves files that aren’t read for a set number of days into cheaper classes, and can move them back on access.
Access points give applications their own root directory and POSIX identity.
Mount with the EFS mount helper (amazon-efs-utils), with TLS for encryption in transit.
Amazon FSx
Type
For
FSx for Windows File Server
SMB shares with Active Directory; Single-AZ or Multi-AZ
FSx for Lustre
HPC and ML; links to S3; scratch (temporary) or persistent deployments
FSx for NetApp ONTAP
NFS, SMB and iSCSI; NetApp features
FSx for OpenZFS
ZFS workloads, very low latency NFS
Troubleshooting mounts
Symptom
Check
Mount times out
Security group on the mount target allows NFS (TCP 2049) from the client; a mount target exists in the client’s AZ; DNS resolution in the VPC
Permission denied
File system policy, IAM authorization, POSIX permissions or access point settings
Slow throughput
Throughput mode; Bursting credits used up
Common traps
EFS for Windows. Use FSx for Windows File Server.
Bursting mode running out of credits. Switch to Elastic throughput.
Missing port 2049 in the mount target’s security group.
A public library stores scanned book pages on Amazon EFS, mounted by several EC2 instances that serve the online archive.
Most pages are viewed a few times in the first month after scanning and then rarely, but any page must still open from the same file path when a reader requests it.
The library wants to lower the storage bill with the LEAST operational effort.
What should a CloudOps engineer do?
Choose one answer.
AA. Set an EFS lifecycle policy to move files to Infrequent Access after 30 days and back to Standard on their first access.
BB. Copy files older than 30 days to an S3 Glacier Flexible Retrieval bucket each month with AWS DataSync, then delete them from EFS.
CC. Change the file system's throughput mode to Bursting so that the library pays only for throughput the archive actually uses.
DD. Move the archive to Amazon FSx for Lustre linked to an S3 bucket, so that older files are released from the file system.
EFS has storage classes for files that are read rarely, and lifecycle management moves files into them after a chosen number of days without access. Files keep their paths, so applications need no change, and the policy can move a file back to Standard when it is read again. Copying data to S3 or another file system adds work and changes how files are accessed.
Want more on EFS, FSx and shared storage? The 5 full-length practice sets have 5 more EFS, FSx and shared storage questions, 325 in all, none of them in the free exam. One payment of $14.99, no subscription. See the practice sets →