Features
High-Performance Parallel I/O
Maximum throughput for HPC workloads with minimal latency and maximum bandwidth.
→Multi-Node Concurrent Access
Multiple instances access the same PFS simultaneously with full read/write capabilities.
→Custom Mount Paths
Specify preferred mount locations to integrate with your existing directory structure.
→Scalable Storage
Start with what you need and resize storage capacity as workload demands grow.
→Resize & Manage
Increase PFS size or delete file systems through simple action controls.
→Monitoring
View attached services and mount information for each connected instance.
→Feature Overview
1. High-Performance Parallel I/O
PFS is optimized for maximum throughput in parallel workloads, delivering superior performance for HPC (High-Performance Computing) applications.
- Benefit: Handle intensive I/O operations with minimal latency and maximum bandwidth.
2. Multi-Node Concurrent Access
Multiple compute instances can access the same PFS simultaneously with full read/write capabilities.
- Flexibility: Share data across distributed HPC (High-Performance Computing) workloads without performance degradation.
3. Custom Mount Paths
Specify your preferred mount location when attaching PFS to Instances.
- Integration: Seamlessly integrate with your existing application directory structure.
4. Scalable Storage
Resize your PFS storage capacity as your workload demands grow.
- Cost-Effective: Start with what you need and scale to meet increasing requirements.
Managing Your PFS
Actions Overview
The Actions button provides management options for each PFS:
- Increase PFS Size - Expand the storage capacity of the existing file system
- Delete File-System - Permanently remove the file system and all stored data
Resizing PFS
The PFS resize feature allows you to increase storage capacity to meet growing performance demands.
How to Resize
Before resizing a PFS, you must UNMOUNT it from all attached Instances to prevent data corruption.
-
Prepare for Resize:
- Unmount the PFS from all attached compute instances
- Ensure no active processes are accessing the file system
- Backup important data as a precaution
-
Access Resize Option:
- Click the Actions button for your PFS
- Select Increase PFS Size from the dropdown
-
Select New Size:
- Use the slider or input field to select your desired size
- Ensure the new size is larger than the current size
- Review the pricing changes
- Click Update to proceed
-
Remount After Resize:
- Once resize completes, remount the PFS to your Instances
- Verify the new capacity is reflected in your system
Deleting PFS
Permanently remove a Parallel File-System when it's no longer needed.
Deleting a PFS will permanently remove all data stored on it. This action cannot be undone. Ensure you have backups of any important data before proceeding.
How to Delete
-
Prepare for Deletion:
- Unmount the PFS from all attached Instances
- Backup any data you want to preserve
- Verify no active workloads depend on this storage
-
Access Delete Option:
- Click the Actions button for your PFS
- Select Delete File-System from the dropdown
-
Confirm Deletion:
- A confirmation dialog will appear
- Review the details and understand the implications
- Check the acknowledgment checkbox
- Click Delete to permanently remove the PFS
Monitoring Your PFS
Attached Services
View which compute instances are currently using your PFS by clicking the Services tab.
- Instance Details: See the names of attached Instances
- Mount Information: View the mount paths for each instance
- Management: Easily identify which workloads depend on this storage
Best Practices
Performance Optimization
- Parallel Access Patterns: Design applications to leverage parallel I/O capabilities
- Optimal Block Sizes: Use larger block sizes for sequential operations
- Network Configuration: Ensure high-bandwidth network connectivity
Cost Management
- Right-Size Storage: Provision capacity based on actual performance needs
- Monitor Usage: Track utilization to optimize storage allocation
- Clean Up Unused Data: Regularly remove temporary files and old datasets
Data Safety
- Regular Backups: Implement automated backup strategies for critical data
- Pre-Resize Backups: Always backup before major resize operations
- Unmount Protocol: Follow proper unmount procedures before resize or delete
- Access Control: Limit PFS access to only the Instances that require it
Best Practices for PFS
Design applications to leverage parallel I/O capabilities with larger block sizes for sequential operations.
Always unmount PFS from all attached instances and backup data before resize operations.
Provision capacity based on actual performance needs and monitor utilization to optimize allocation.
Limit PFS access to only the instances that require it and regularly clean up temporary files.