In UNIX systems, mounting filesystems can be done using either hard or soft options, each with distinct behaviors that impact system reliability and performance. The fundamental difference lies in how they handle server failures or network disruptions.

Hard mounts provide stronger consistency guarantees:

# Example hard mount in /etc/fstab
server:/export/data /mnt/data nfs rw,hard,intr 0 0

• Advantages:
  - Guarantees data integrity by retrying indefinitely
  - Better for critical operations where data consistency is paramount
  - Preferred for database storage or transactional systems
• Disadvantages:
  - Can cause processes to hang indefinitely during network partitions
  - May require manual intervention to recover frozen processes

Soft mounts offer more flexibility but weaker guarantees:

# Example soft mount configuration
mount -t nfs -o soft,timeo=30,retrans=3 server:/home /mnt/home

• Advantages:
  - Prevents indefinite hangs by failing after retries
  - Better for interactive systems where responsiveness is critical
  - Suitable for non-critical data like user home directories
• Disadvantages:
  - Potential data corruption if writes fail during timeout
  - Higher risk of silent failures going unnoticed

Benchmarking shows significant differences:

# Simple IOPS test comparison
dd if=/dev/zero of=/mnt/hard/testfile bs=1M count=1000
dd if=/dev/zero of=/mnt/soft/testfile bs=1M count=1000

Hard mounts typically show 15-20% lower throughput during network instability but more consistent performance under normal conditions.

• Use hard mounts for:
  - Database filesystems (Oracle, PostgreSQL)
  - Financial transaction systems
  - Any application where data integrity > availability
• Consider soft mounts for:
  - User home directories
  - Development environments
  - Read-heavy workloads where stale data is acceptable

For hung NFS processes on hard mounts:

# List processes waiting on NFS
lsof | grep -i nfs | grep -i DEL

# Alternative using /proc
find /proc/*/fd -ls | grep -i nfs

---

In UNIX systems, mounting filesystems can be done using either hard or soft options, each with distinct behaviors:

# Hard mount example
mount -o hard server:/share /mnt/nfs

# Soft mount example
mount -o soft,timeo=30 server:/share /mnt/nfs

Advantages:

• Guarantees data consistency - retries indefinitely until successful
• Better for critical operations where data integrity is paramount
• Default behavior for many enterprise storage systems

Disadvantages:

• Can cause processes to hang indefinitely during network outages
• May lead to system unresponsiveness if server becomes unavailable
• Not suitable for unreliable networks

Advantages:

• Fails after specified retries (configurable via timeo parameter)
• Prevents application hangs during temporary network issues
• Better for non-critical data or read-only operations

Disadvantages:

• Potential data corruption if writes fail silently
• Not recommended for database operations
• May require additional error handling in applications

Benchmarking shows significant differences in timeout behavior:

# Measure timeout impact
time dd if=/mnt/nfs/largefile of=/dev/null bs=1M

Hard mounts typically show more consistent throughput but can exhibit severe latency spikes during network issues.

For mission-critical systems:

mount -o hard,intr,noatime,nodev,nosuid server:/critical /mnt/important

For temporary or development environments:

mount -o soft,timeo=60,retrans=3 server:/dev /mnt/temp

• Use hard mounts for financial systems, databases, and transaction logs
• Consider soft mounts for web content, media files, and temporary data
• Always monitor NFS stats with nfsstat -m
• Combine with HA solutions for critical mounts

When experiencing mount issues:

# Check current mounts
mount | grep nfs

# View NFS statistics
cat /proc/fs/nfsfs/servers