sstablelevelreset

Resets all SSTables to level 0 for tables using Leveled Compaction Strategy (LCS).

Synopsis

sstablelevelreset --really-reset <keyspace> <table>

Description

sstablelevelreset resets the compaction level of all SSTables for a table back to level 0 (L0). This is specifically for tables using Leveled Compaction Strategy (LCS) where level corruption or distribution issues have occurred.

This tool is used when:

Level metadata is corrupted - SSTables have incorrect level assignments
Compaction is stuck - Levels are imbalanced and compaction cannot progress
After switching to LCS - Existing SSTables need level assignment
Recovery from issues - Compaction problems require a fresh start

Cassandra Must Be Stopped

Cassandra must be completely stopped before running sstablelevelreset. Running this tool while Cassandra is active will cause data corruption.

How LCS Levels Work

Level Properties

Level	Max Size	Non-overlapping	SSTables
L0	Up to 4 × `sstable_size_in_mb`	No	1-4
L1	`fanout_size` × L0 size	Yes	~10
L2	`fanout_size` × L1 size	Yes	~100
L3+	`fanout_size` × previous	Yes	~1000+

Configuration-Dependent Values

Level sizes depend on LCS compaction strategy settings:

sstable_size_in_mb: Target SSTable size (default: 160 MiB)
fanout_size: Size multiplier between levels (default: 10)

With defaults, L0 holds up to ~640 MiB (4 × 160 MiB) before triggering compaction to L1.

Arguments

Argument	Description
`keyspace`	Name of the keyspace containing the table
`table`	Name of the table to reset levels

Options

Option	Description
`--really-reset`	Required safety flag to confirm reset

The --really-reset flag is mandatory to prevent accidental execution.

Examples

Basic Usage

# Stop Cassandra first
sudo systemctl stop cassandra

# Reset all SSTables to level 0
sstablelevelreset --really-reset my_keyspace my_table

# Start Cassandra - LCS will reorganize levels
sudo systemctl start cassandra

Verify Current Levels

# Before reset - check level distribution
sstablemetadata /var/lib/cassandra/data/my_keyspace/my_table-*/*-Data.db | grep "SSTable Level"

Reset All LCS Tables

#!/bin/bash
# reset_lcs_tables.sh - Reset levels for all LCS tables

KEYSPACE="$1"

# Find tables using LCS
lcs_tables=$(cqlsh -e "SELECT table_name FROM system_schema.tables
    WHERE keyspace_name = '$KEYSPACE'" | \
    while read table; do
        compaction=$(cqlsh -e "SELECT compaction FROM system_schema.tables
            WHERE keyspace_name = '$KEYSPACE' AND table_name = '$table'" 2>/dev/null | \
            grep -i "leveledcompaction")
        if [ -n "$compaction" ]; then
            echo "$table"
        fi
    done)

for table in $lcs_tables; do
    echo "Resetting levels for ${KEYSPACE}.${table}..."
    sstablelevelreset --really-reset "$KEYSPACE" "$table"
done

When to Use sstablelevelreset

Scenario 1: Corrupted Level Metadata

# Symptoms:
# - Cassandra logs show "Level X has Y SSTables but should have Z"
# - Unexpected compaction behavior
# - Compaction stats show wrong level distribution

# Check current state
nodetool cfstats my_keyspace.my_table | grep -i level

# Reset levels
sudo systemctl stop cassandra
sstablelevelreset --really-reset my_keyspace my_table
sudo systemctl start cassandra

Scenario 2: After Switching to LCS

# When migrating from STCS to LCS

# 1. Change compaction strategy
cqlsh -e "ALTER TABLE my_keyspace.my_table WITH compaction = {
    'class': 'LeveledCompactionStrategy'
};"

# 2. Stop Cassandra
sudo systemctl stop cassandra

# 3. Reset levels (existing SSTables may have level 0 already, but reset ensures clean state)
sstablelevelreset --really-reset my_keyspace my_table

# 4. Start Cassandra - LCS will organize SSTables into levels
sudo systemctl start cassandra

# 5. Monitor compaction progress
nodetool compactionstats

Scenario 3: Compaction Stuck

# Symptoms:
# - Compaction not progressing
# - High number of SSTables
# - Levels are imbalanced

# Check level distribution
for f in /var/lib/cassandra/data/my_keyspace/my_table-*/*-Data.db; do
    level=$(sstablemetadata "$f" | grep "SSTable Level:" | awk '{print $3}')
    echo "L${level}: $(basename $f)"
done | sort | uniq -c

# If distribution is problematic, reset
sudo systemctl stop cassandra
sstablelevelreset --really-reset my_keyspace my_table
sudo systemctl start cassandra

Scenario 4: Recovery from sstableofflinerelevel Issues

# If sstableofflinerelevel caused problems
# Reset to clean state and let LCS reorganize

sudo systemctl stop cassandra
sstablelevelreset --really-reset my_keyspace my_table
sudo systemctl start cassandra

Impact of Level Reset

Before Reset

Level Distribution:
L0: 2 SSTables (320 MB)
L1: 10 SSTables (1.6 GB)
L2: 87 SSTables (14 GB)
L3: 432 SSTables (69 GB)

After Reset

Level Distribution:
L0: 531 SSTables (85 GB)  <- All SSTables moved here
L1: 0 SSTables
L2: 0 SSTables
L3: 0 SSTables

After Compaction Reorganization

Level Distribution:
L0: 4 SSTables (640 MB)   <- Back to normal
L1: 10 SSTables (1.6 GB)
L2: 100 SSTables (16 GB)
L3: 450 SSTables (72 GB)

Post-Reset Actions

1. Start Cassandra

sudo systemctl start cassandra

2. Monitor Compaction

# Watch compaction progress
watch -n 5 'nodetool compactionstats'

# Check level distribution
nodetool cfstats my_keyspace.my_table | grep -A10 "SSTables in each level"

3. Verify Health

# Ensure reads work correctly
cqlsh -e "SELECT * FROM my_keyspace.my_table LIMIT 10;"

# Check for compaction errors
grep -i "error\|exception" /var/log/cassandra/system.log