nodetool clearsnapshot

Removes one or more snapshots from the node to reclaim disk space.

---

Synopsis

nodetool [connection_options] clearsnapshot [options] [--] [keyspace ...]

See connection options for connection options.

Description

nodetool clearsnapshot deletes snapshot files from the local node. Since snapshots are hard links to SSTable files, they consume disk space when the original SSTables are compacted away. Regular cleanup of old snapshots is essential for disk space management.

---

Arguments

Argument	Description
keyspace	Keyspace(s) to clear snapshots from. If omitted, clears from all keyspaces

---

Options

Option	Description
-t, --tag	Snapshot tag to clear. Required unless using --all
--all	Clear all snapshots

---

Examples

Clear Specific Snapshot

nodetool clearsnapshot -t my_backup

Removes the snapshot tagged my_backup from all keyspaces.

Clear Snapshot from Specific Keyspace

nodetool clearsnapshot -t my_backup my_keyspace

Clear All Snapshots

nodetool clearsnapshot --all

Use with Caution

--all removes every snapshot on the node. Verify no critical backups exist before running.

Clear All Snapshots from Keyspace

nodetool clearsnapshot --all my_keyspace

---

When to Use

After Successful Backup

Once backup files have been copied off-node:

# After copying snapshot files to backup storage
nodetool clearsnapshot -t backup_20240115

Reclaim Disk Space

When disk usage is high due to old snapshots:

# Check snapshot sizes first
nodetool listsnapshots

# Remove old snapshots
nodetool clearsnapshot -t old_backup

Clean Up After Testing

Remove snapshots created during testing:

nodetool clearsnapshot -t test_snapshot

Regular Maintenance

Scheduled cleanup of aged snapshots:

# Find snapshots older than 7 days and clear them
# (Requires scripting based on listsnapshots output)
nodetool clearsnapshot -t weekly_backup_$(date -d "7 days ago" +%Y%m%d)

---

When NOT to Use

Before Verifying Backup

Verify Before Clearing

Never clear snapshots until backup files are:

• Copied to remote storage
• Verified for integrity
• Tested for recoverability

On Active Snapshots

Don't clear snapshots that are currently being copied or used.

Without Checking First

Always list snapshots before clearing:

nodetool listsnapshots

---

Verification Before Clearing

List All Snapshots

nodetool listsnapshots

Output:

Snapshot name    Keyspace name    Column family name    True size    Size on disk
backup_20240115  my_keyspace      users                 1.5 GB       1.5 GB
backup_20240115  my_keyspace      orders                2.3 GB       2.3 GB
backup_20240108  my_keyspace      users                 1.2 GB       0 bytes

Check Specific Snapshot

nodetool listsnapshots | grep backup_20240115

Verify Disk Usage

# Check total snapshot space
du -sh /var/lib/cassandra/data/*/*/snapshots/

# Check specific snapshot
du -sh /var/lib/cassandra/data/*/*/snapshots/backup_20240115/

---

Space Reclamation

How Space is Freed

State	SSTable	Snapshot	Disk Usage
Before clearsnapshot	Active (new)	Hard link exists	Space shared via inode
After clearsnapshot	Active (new)	Deleted	Space freed if no other references

Space Reclamation

• Hard links are removed by clearsnapshot
• Disk blocks are freed only if no other references exist
• Space reclamation may not be immediate depending on filesystem

Verifying Space Freed

# Before
df -h /var/lib/cassandra

# Clear snapshot
nodetool clearsnapshot -t old_backup

# After
df -h /var/lib/cassandra

---

Automated Cleanup

Script for Aged Snapshots

#!/bin/bash
# Clear snapshots older than specified days

DAYS_OLD=7
KEYSPACE="my_keyspace"

# List snapshots and filter by age
nodetool listsnapshots | while read line; do
    # Parse snapshot name and check date
    # Implementation depends on naming convention
done

Using TTL Snapshots (4.0+)

Instead of manual cleanup, create snapshots with TTL:

# Snapshot auto-deletes after 24 hours
nodetool snapshot -t temp_backup --ttl 24h my_keyspace

---

Common Issues

"Snapshot does not exist"

ERROR: Snapshot 'my_backup' does not exist

The snapshot tag wasn't found:

# List actual snapshot names
nodetool listsnapshots

Space Not Freed

If disk space doesn't decrease after clearing:

1. Check if files were actually removed:

   ls /var/lib/cassandra/data/*/*/snapshots/

2. Hard links may still exist (original SSTable not yet compacted)

3. Filesystem may not report freed space immediately

Cannot Delete Snapshot Directory

Permission issues:

# Check ownership
ls -la /var/lib/cassandra/data/my_keyspace/users-*/snapshots/

# Cassandra user should own files
sudo chown -R cassandra:cassandra /var/lib/cassandra/data/

---

Best Practices

Snapshot Cleanup Guidelines

1. List before clearing - Verify which snapshots exist
2. Use specific tags - Avoid --all in production
3. Verify backups first - Ensure data is safely stored elsewhere
4. Automate cleanup - Use TTL or scheduled scripts
5. Monitor disk usage - Track snapshot growth over time
6. Document retention - Define snapshot retention policies

Retention Policy Example

Snapshot Type	Retention
Pre-maintenance	24 hours after maintenance
Daily backup	7 days
Weekly backup	30 days
Monthly backup	1 year

---

Related Commands

Command	Relationship
snapshot	Create snapshots
listsnapshots	List existing snapshots
tablestats	Check snapshot space per table