Voltage Park: mounting shared storage

Mount persistent storage volumes on Voltage Park instances using Network File System (NFS) protocol.

Overview

Voltage Park persistent volumes use NFS (Network File System) version 3 protocol for high-performance network-attached storage. Each volume is accessible via a virtual IP address and supports concurrent access from multiple instances.

Protocol: NFS v3 Connection: Virtual IP-based Performance: Parallel connections via nconnect option

Prerequisites

Before starting, ensure you have:

1. Created a volume with provider: voltage-park via the Spheron dashboard (Create Volume page)
2. Attached the volume to your Voltage Park instance
3. Retrieved the volume's virtual IP address from:
   • Instance Details drawer → "Mounting your storage volume" section, or
   • Volumes page
4. SSH access to your Voltage Park instance

Mounting process

ssh ubuntu@<your-instance-ip>

sudo apt install nfs-common

sudo mkdir /data

echo '<virtualIP>:/data /data nfs rw,nconnect=16,nfsvers=3 0 0' | sudo tee -a /etc/fstab

echo '192.168.100.50:/data /data nfs rw,nconnect=16,nfsvers=3 0 0' | sudo tee -a /etc/fstab

sudo mount -a

df -h

Filesystem              Size  Used Avail Use% Mounted on
...
192.168.100.50:/data    100G   1.0G   99G   1% /data

sudo chown ubuntu:ubuntu /data

Use your mounted volume

Access the storage

Once mounted, use the storage like any local directory:

# Navigate to the storage
cd /data
 
# Create files and directories
mkdir my-project
echo "Hello, storage!" > my-project/readme.txt
 
# List contents
ls -lh /data/

Check storage usage

Monitor your volume's space usage:

# Check space on the mounted volume
df -h /data
 
# Check detailed disk usage
du -sh /data/*

Work with large datasets

The mounted volume is ideal for:

• ML/AI training datasets
• Model checkpoints and artifacts
• Shared data across multiple instances
• Persistent application data

# Example: Download dataset to storage
cd /data
wget https://example.com/large-dataset.tar.gz
tar -xzf large-dataset.tar.gz

Unmount shared storage

Unmount shared storage when swapping out the attached storage volume, detaching the volume from the instance, or terminating the instance.

Unmount the volume

sudo umount /data

If you get a "target is busy" error, ensure no processes are using the storage:

# Check what's using the storage
lsof /data
 
# Or use fuser
fuser -m /data

Remove from file system table

Remove the volume configuration from /etc/fstab to prevent auto-mount on next boot:

# Edit fstab and remove the NFS entry
sudo nano /etc/fstab
 
# Or use sed to remove it automatically
sudo sed -i '/\/data.*nfs/d' /etc/fstab

Troubleshooting

Mount fails with "Connection refused"

Cause: Incorrect virtual IP or volume not attached to instance.

1. Verify the volume is attached to your instance via the dashboard or API
2. Double-check the virtual IP address
3. Ensure the virtual IP in /etc/fstab matches the volume's virtual IP

Mount fails with "No such file or directory"

Cause: Mount point directory doesn't exist.

sudo mkdir -p /data
sudo mount -a

Volume not mounting on boot

Cause: Network not ready when fstab mounts are processed.

Solution: Add the _netdev option to the fstab entry:

<storage-vip>:/data /data nfs rw,nconnect=16,nfsvers=3,_netdev 0 0

The _netdev option tells the system to wait for the network before mounting.

Performance issues

Cause: Suboptimal NFS settings or network latency.

• Increase the nconnect value for more parallel connections (try 32 or 64)
• Use NFSv4 if supported: nfsvers=4
• Add async mode for better write performance: async

<storage-vip>:/data /data nfs rw,nconnect=32,nfsvers=4,async 0 0

Check mount status

# List all NFS mounts
mount | grep nfs
 
# Check NFS statistics
nfsstat
 
# Verify fstab syntax
sudo mount -fav

Best practices

• Use descriptive mount points: /data, /models, /datasets
• Create subdirectories for different projects or datasets
• Document what data is stored where

• Use higher nconnect values (16 to 64) for better throughput
• Consider async mode for write-heavy workloads
• Monitor network bandwidth usage

• Maintain backups of critical data
• Volumes persist, but data can still be lost due to corruption or accidental deletion
• Test backup and restore procedures

• Restrict access to the mount point using filesystem permissions
• Only mount volumes from trusted sources
• Regularly audit who has access to shared storage

• Add mount commands to startup scripts for new instances
• Use cloud-init to configure NFS mounts automatically

Example cloud-init script:

#cloud-config
runcmd:
  - apt-get install -y nfs-common
  - mkdir -p /data
  - echo '192.168.100.50:/data /data nfs rw,nconnect=16,nfsvers=3 0 0' >> /etc/fstab
  - mount -a
  - chown ubuntu:ubuntu /data

Mount multiple volumes

You can attach and mount multiple storage volumes to a single Voltage Park instance:

# Create mount points
sudo mkdir -p /data1 /data2 /models
 
# Add to fstab
echo '192.168.100.50:/data /data1 nfs rw,nconnect=16,nfsvers=3 0 0' | sudo tee -a /etc/fstab
echo '192.168.100.51:/data /data2 nfs rw,nconnect=16,nfsvers=3 0 0' | sudo tee -a /etc/fstab
echo '192.168.100.52:/data /models nfs rw,nconnect=16,nfsvers=3 0 0' | sudo tee -a /etc/fstab
 
# Mount all
sudo mount -a

Share volumes across instances

To share a volume across instances simultaneously:

1. Attach the same volume to multiple instances via the API or dashboard
2. Mount the volume on each instance using the same virtual IP
3. All instances can read and write to the shared storage

What's next

• Volume Mounting Overview: Multi-provider volume mounting guide
• Volume API Reference: Create and manage volumes programmatically
• External Storage Access: Configure local NVMe drives on Voltage Park
• SSH Connection Setup: Secure instance access
• Security Best Practices: Secure your storage and data