nodetool disablebackup

Disables incremental backup on the node.

---

Synopsis

nodetool [connection_options] disablebackup

See connection options for connection options.

Description

nodetool disablebackup disables incremental backup mode on the node. Once disabled, Cassandra stops creating hard links to newly flushed SSTables. Existing backup files in the backups directory are not removed—they must be cleaned up manually.

Incremental backup creates hard links to new SSTables when they are created (during memtable flush, compaction, and other SSTable-producing operations), providing continuous backup capability. Disabling this feature stops the automatic link creation process.

Non-Persistent Setting

This setting is applied at runtime only and does not persist across node restarts. After a restart, incremental backup reverts to the incremental_backups setting in cassandra.yaml (default: false).

To make the change permanent, update cassandra.yaml:

incremental_backups: false

---

Behavior

When incremental backup is disabled:

1. No new hard links are created for flushed SSTables
2. Existing backup files remain in place (not deleted)
3. Disk space used by existing backups is retained
4. The change takes effect immediately
5. The setting reverts to the cassandra.yaml configuration on restart

---

Examples

Basic Usage

nodetool disablebackup

Verify Disabled

nodetool disablebackup
nodetool statusbackup
# Expected output: not running

---

When to Use

Disk Space Management

When backup files consume too much space:

# Check current backup sizes
du -sh /var/lib/cassandra/data/*/*/backups/
# Total: 150GB

# Stop new backups
nodetool disablebackup

# Copy existing backups to external storage
rsync -av /var/lib/cassandra/data/*/*/backups/ /backup/server/

# Clean up local backups
find /var/lib/cassandra/data -path "*/backups/*" -type f -delete

# Verify cleanup
du -sh /var/lib/cassandra/data/*/*/backups/

Temporary Maintenance Disable

During maintenance operations:

# Disable during heavy operations
nodetool disablebackup

# Perform maintenance (repair, compaction, etc.)
nodetool repair -pr my_keyspace

# Re-enable after maintenance
nodetool enablebackup

# Verify
nodetool statusbackup

Switching Backup Strategy

When migrating to a different backup solution:

# Disable incremental backup
nodetool disablebackup

# Clean up existing backup files
find /var/lib/cassandra/data -path "*/backups/*" -delete

# (Configure new backup solution)

I/O Reduction

During I/O-intensive operations:

# Disable to reduce I/O overhead
nodetool disablebackup

# Perform I/O-intensive operation
nodetool upgradesstables

# Re-enable
nodetool enablebackup

Before Node Decommission

Remove unnecessary backup overhead before decommission:

# Disable backups (node is leaving)
nodetool disablebackup

# Clean up backup files
find /var/lib/cassandra/data -path "*/backups/*" -delete

# Proceed with decommission
nodetool decommission

---

Impact Assessment

What Changes

Aspect	Before Disable	After Disable
New backup links	Created on flush	Not created
Existing backups	Present	Still present
Disk overhead	Grows with flushes	No new growth
Recovery capability	Full incremental	Depends on existing files

What Does NOT Change

• Existing backup files remain (must clean manually)
• Snapshots are unaffected
• Compaction behavior unchanged
• Normal data operations continue

---

Cleanup After Disable

Disabling backup does not remove existing files:

Check Backup Size

# Per-keyspace backup sizes
du -sh /var/lib/cassandra/data/*/*/backups/

# Total backup size
find /var/lib/cassandra/data -path "*/backups/*" -type f -exec du -ch {} + | tail -1

Copy Before Cleanup

# Back up to external storage first
rsync -av /var/lib/cassandra/data/*/*/backups/ /backup/final_incremental/

Remove Backup Files

# Remove all backup files (careful!)
find /var/lib/cassandra/data -path "*/backups/*" -type f -delete

# Or remove per-keyspace
rm -rf /var/lib/cassandra/data/my_keyspace/*/backups/*

# Verify removal
find /var/lib/cassandra/data -path "*/backups/*" -type f | wc -l
# Should be 0

---

Workflow: Complete Backup Disable and Cleanup

disable_and_cleanup_backup.sh

#!/bin/bash
echo "=== Disable Incremental Backup ==="
echo ""

# 1. Check current state
echo "1. Current status: $(nodetool statusbackup)"

# 2. Check current backup size
echo "2. Current backup size:"
find /var/lib/cassandra/data -path "*/backups/*" -type f -exec du -ch {} + 2>/dev/null | tail -1

# 3. Disable backup
echo "3. Disabling incremental backup..."
nodetool disablebackup

# 4. Verify disabled
echo "4. New status: $(nodetool statusbackup)"

# 5. Optional: Clean up existing backups
read -p "5. Remove existing backup files? (y/n): " cleanup
if [ "$cleanup" = "y" ]; then
    echo "   Removing backup files..."
    find /var/lib/cassandra/data -path "*/backups/*" -type f -delete
    echo "   Done."
fi

echo ""
echo "=== Complete ==="

---

Effects on Disk Space

Hard Links Explained

When backup is enabled, SSTables have hard links in the backups directory:

SSTable: /data/keyspace/table/nb-1-big-Data.db
Backup:  /data/keyspace/table/backups/nb-1-big-Data.db
         (both point to same disk blocks)

After compaction removes the original SSTable:

• The backup file becomes the only reference
• Disk space is now exclusively used by backup

Space Recovery

To reclaim space after disabling:

# Check space used by backups
du -sh /var/lib/cassandra/data/*/*/backups/

# Remove to reclaim
find /var/lib/cassandra/data -path "*/backups/*" -delete

# Verify space recovered
df -h /var/lib/cassandra

---

Runtime vs Persistent Configuration

Runtime (Temporary)

nodetool disablebackup
# Does not persist across restart

Persistent (Permanent)

Edit cassandra.yaml:

incremental_backups: false

Then either restart or use nodetool:

# For immediate effect without restart
nodetool disablebackup

Comparison

Method	Effect	Persistence
nodetool disablebackup	Immediate	Until restart
cassandra.yaml: false	After restart	Permanent
Both	Immediate + permanent	Best practice

---

Monitoring

Verify Status

nodetool statusbackup
# Expected: not running

Monitor Backup Directory Growth

Even after disable, verify no new files:

# Before operation
ls /var/lib/cassandra/data/my_keyspace/my_table-*/backups/ | wc -l
# Count: 50

# Force flush
nodetool flush my_keyspace

# After flush (should be same if disabled)
ls /var/lib/cassandra/data/my_keyspace/my_table-*/backups/ | wc -l
# Count: 50 (unchanged)

---

Troubleshooting

Status Still Shows "running"

# Retry disable
nodetool disablebackup

# Check JMX connectivity
nodetool info

# Verify
nodetool statusbackup

Backup Files Still Appearing

If new backup files appear after disable:

# Verify disabled
nodetool statusbackup
# Should be: not running

# Check cassandra.yaml (might override on restart)
grep incremental_backups /etc/cassandra/cassandra.yaml

# If yaml says true, update it
vim /etc/cassandra/cassandra.yaml
# Set: incremental_backups: false

Cannot Remove Backup Files

If deletion fails:

# Check file ownership
ls -la /var/lib/cassandra/data/my_keyspace/*/backups/

# Check if files are open
lsof +D /var/lib/cassandra/data/my_keyspace/my_table/backups/

# May need to run as cassandra user
sudo -u cassandra rm -rf /var/lib/cassandra/data/*/*/backups/*

---

Cluster-Wide Disable

Disable backup across all nodes:

disable_backup_cluster.sh

#!/bin/bash
# Get list of node IPs from local nodetool status
nodes=$(nodetool status | grep "^UN" | awk '{print $2}')

echo "Disabling incremental backup cluster-wide"
echo "==========================================="

for node in $nodes; do
    echo -n "$node: "
    ssh "$node" "nodetool disablebackup 2>/dev/null && echo 'disabled' || echo 'failed'"
done

echo ""
echo "Verification:"
for node in $nodes; do
    echo -n "$node: "
    ssh "$node" "nodetool statusbackup"
done

---

Best Practices

Disable Backup Guidelines

1. Document the reason - Note why backup was disabled
2. Copy before cleanup - Always backup before deleting files
3. Update cassandra.yaml - Make change persistent if intended
4. Verify status - Confirm disabled with statusbackup
5. Clean up disk - Remove old backup files to reclaim space
6. Cluster consistency - Apply to all nodes if disabling cluster-wide
7. Re-enable after maintenance - Don't forget to restore backup if temporary

---

Related Commands

Command	Relationship
enablebackup	Enable incremental backup
statusbackup	Check backup status
snapshot	Full point-in-time backup
listsnapshots	List existing snapshots
clearsnapshot	Remove snapshots