Cloud concept · Updated

Shared File Systems - Shared file systems provide a common hierarchy of files and directories that multiple authorized clients can access over a network.

In short

A shared file system lets applications on different servers work with the same files through familiar file operations rather than copying data between machines. In the cloud, managed file services support shared application content, user directories, and processing workloads while reducing the work of operating storage servers.

Cloud Concepts Hub →

6 key aspects of Shared File Systems

Shared file systems combine common file access with protocol compatibility, concurrency controls, persistent storage, workload-specific performance, and layered protection.

01
Common file namespace

Multiple clients access the same directory hierarchy instead of maintaining separate copies. This supports shared uploads, user home directories, and datasets used by distributed workers.

02
Protocol and application compatibility

Clients connect through supported protocols such as NFS or SMB. Match the application's requirements for permissions, locking, identity integration, and file operations to the chosen filesystem.

03
Coordinated concurrent access

Many clients can access a shared filesystem, but conflicting updates still require coordination. Validate locking and caching behavior before allowing multiple application instances to modify the same files.

04
Independent storage lifecycle

Shared files can persist independently of individual compute instances. Replacement servers can reconnect to the same storage, provided their network access and permissions are configured correctly.

05
Workload-dependent performance

Latency, throughput, IOPS, and metadata performance all influence application behavior. Small-file workloads and large sequential transfers can require different performance configurations.

06
Layered security and recovery

Control both network connectivity and access to individual files or directories. Select suitable availability options and maintain recovery mechanisms because replication alone does not protect against deletion or corruption.

Shared File Systems exam questions

Real practice exam questions on Shared File Systems, with the correct answer and its explanation. Answer them with every option in the exam simulator.

CLF-C02

A company is moving a Windows application to multiple Amazon EC2 instances. All instances must access the same shared folders over SMB, use Microsoft Active Directory authentication, and preserve native Windows NTFS permissions. The company wants AWS to manage the file server infrastructure. Which solution meets these requirements?

Show answerCORRECT ANSWER

Create an Amazon FSx for Windows File Server file system and connect every application instance to its shared folders.

Amazon FSx for Windows File Server provides fully managed shared storage with SMB support, Microsoft Active Directory integration, and native Windows NTFS permissions. Multiple application instances can access the same files without operating their own file server.

SOA-C03

A company runs a Linux analytics cluster on Amazon EC2 instances that share an SSD-based Amazon FSx for Lustre file system. Many workers concurrently read different sections of one large reference file. Investigation shows that the file has a stripe count of 1 and that its object storage target (OST) is saturated, while other OSTs are mostly idle. The clients have sufficient CPU and network capacity.

A CloudOps engineer must distribute reads of the existing reference file across multiple OSTs without increasing storage capacity or provisioned throughput. A maintenance window is available to replace the file.

Which action should the engineer take?

Show answerCORRECT ANSWER

Create a replacement file with a stripe count that spans multiple OSTs, copy the reference data into it, and update the workers to use the replacement file.

Lustre file striping distributes a file's contents across OSTs. Creating the replacement with a higher stripe count before copying the data spreads the file's I/O across multiple targets, allowing concurrent readers to use otherwise idle storage resources.

DOP-C02

A company uses AWS CodeDeploy to perform in-place deployments to Linux EC2 instances in an Auto Scaling group. All instances mount the same Amazon EFS directory at /opt/orders/uploads. The application stores customer uploads there, and its upload path is configurable. Application binaries are installed under /opt/orders/releases/<revision>.

A deployment hook synchronizes the release bundle into /opt/orders by using rsync --delete. The bundle does not contain customer uploads. During a deployment to one instance, customer files disappear from every instance. Investigation confirms that the synchronization traversed the uploads mount and deleted files that were absent from the bundle.

The company must continue accepting uploads throughout deployments. Application rollback must restore the previous binaries without reverting customer data. Instances must remain disposable, and all instances must continue accessing the same uploads. The company wants to retain its current storage service and deployment mechanism.

Which change should a DevOps engineer implement?

Show answerCORRECT ANSWER

Mount the shared uploads directory outside /opt/orders and configure the application to use that path. Limit deployment synchronization and cleanup to instance-local release directories, and make rollback select the previous local release without modifying the shared uploads.

A deletion through an EFS mount changes the shared filesystem, so it affects every client. Separating persistent shared data from deployment-managed files prevents release cleanup from deleting uploads. Each instance can replace or roll back its local binaries independently while continuing to use the same current customer data.

Practice Shared File Systems in a full exam

Questions on Shared File Systems in each practice exam, with every answer option and an explanation for each:

Start a free practice exam

What is Shared File Systems

2 min read

A shared file system presents files and directories through a common namespace that multiple clients can access over a network. Applications can open, read, write, rename, and delete files using supported operating-system interfaces. This makes shared file storage useful when applications expect filesystem paths and need access to the same persistent data from more than one machine.

Clients usually mount a network filesystem or connect to a network share using a protocol such as Network File System (NFS) or Server Message Block (SMB). NFS is common in Linux and Unix environments, while SMB is common in Windows environments. Protocol support, permissions, locking behavior, and filesystem semantics must match the application. Supporting a file protocol does not automatically make a service compatible with every application that uses that protocol.

In the cloud, a managed file service operates the underlying storage infrastructure while engineers configure network access, client mounts, permissions, capacity or performance settings, and data protection. For example, several web servers can share uploaded images so that replacing an individual server does not remove those images. A Windows organization might provide shared departmental folders using SMB and directory-based authentication. A compute cluster might use a parallel shared filesystem to process a large dataset from many workers.

For AWS service identification, Amazon Elastic File System (Amazon EFS) provides managed NFS file storage commonly used with Linux workloads. Amazon FSx is a family of managed services for specific filesystems, including Windows File Server, Lustre, NetApp ONTAP, and OpenZFS. The practical selection question is which protocol, filesystem capabilities, performance characteristics, and availability options the workload requires. Not every Amazon FSx option supports the same clients or provides the same features.

Shared file storage differs from both block and object storage. Block storage presents a disk-like device on which an operating system typically creates a filesystem; attaching a block device to several machines does not by itself create safe shared file access. Object storage exposes objects through APIs rather than providing the same semantics as a conventional network filesystem. A bucket is therefore not automatically a drop-in replacement for an application that depends on file locking, directory operations, or in-place updates.

Sharing access also does not make concurrent writes automatically safe. Applications must coordinate updates through supported locking mechanisms or application-level controls. Network latency, metadata-heavy operations, throughput limits, and client caching can affect performance. Engineers should test representative workloads rather than judge a service solely by its storage capacity.

Finally, availability, durability, and backup are separate concerns. A deployment that spans availability zones can improve resilience, but not every file service or deployment option does so. Redundancy does not prevent authorized deletion, ransomware, or application errors. Restrict network access, enforce appropriate file permissions, use supported encryption controls, and maintain tested backups or recovery mechanisms. Include storage, performance, backup, and applicable data-transfer charges when comparing costs.

Learn Shared File Systems in lessons

10 short lessons on Shared File Systems in 5 certification courses.

and 4 more lessons

Open the lessons