Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
Point-in-time snapshots of the dataset. Compact binary format.
Configuration
# redis.conf
# Auto-save triggers (any condition triggers save)
save 900 1 # Save after 900s if ≥1 key changed
save 300 10 # Save after 300s if ≥10 keys changed
save 60 10000 # Save after 60s if ≥10000 keys changed
# Disable RDB completely
# save ""
# File settings
dbfilename dump.rdb
dir /var/lib/redis
# Compression (uses LZF)
rdbcompression yes
# Checksum (uses CRC64)
rdbchecksum yes
# Error handling
stop-writes-on-bgsave-error yes
Commands
SAVE # Blocking save (AVOID in production)
BGSAVE # Background save (fork)
LASTSAVE # Last successful save timestamp
# Check status
INFO persistence
# rdb_bgsave_in_progress:0
# rdb_last_save_time:1704067200
# rdb_last_bgsave_status:ok
# rdb_changes_since_last_save:0
RDB Advantages
Compact single-file backup
Fast recovery (just load file)
Good for disaster recovery
Minimal performance impact
RDB Disadvantages
Data loss between snapshots
Fork can be slow with large datasets
Not suitable for zero data loss requirements
AOF (Append-Only File)
Write-ahead log recording every write operation.
Configuration
# redis.conf
# Enable AOF
appendonly yes
appendfilename "appendonly.aof"
appenddirname "appendonlydir"
# Fsync policies
# always - Fsync after every write (safest, slowest)
# everysec - Fsync every second (recommended)
# no - Let OS decide (fastest, least safe)
appendfsync everysec
# Rewrite settings
auto-aof-rewrite-percentage 100 # Rewrite if AOF is 2x size since last rewrite
auto-aof-rewrite-min-size 64mb # Minimum size to trigger rewrite
# Don't fsync during rewrite (better performance)
no-appendfsync-on-rewrite no
# Truncate corrupted AOF
aof-load-truncated yes
# Enable RDB preamble in AOF
aof-use-rdb-preamble yes
Commands
BGREWRITEAOF # Compact AOF file
# Check status
INFO persistence
# aof_enabled:1
# aof_rewrite_in_progress:0
# aof_current_size:123456
# aof_base_size:100000
#!/bin/bash# recover-redis.sh# 1. Stop Redissudo systemctl stop redis
# 2. Backup current (corrupted) datamv /var/lib/redis /var/lib/redis.corrupted
# 3. Restore from backupmkdir -p /var/lib/redis
cp /backup/redis/dump_YYYYMMDD.rdb /var/lib/redis/dump.rdb
# Or for AOF:
tar -xzf /backup/redis/aof_YYYYMMDD.tar.gz -C /var/lib/redis/
# 4. Fix permissionschown -R redis:redis /var/lib/redis
# 5. Start Redissudo systemctl start redis
# 6. Verify
redis-cli PING
redis-cli INFO keyspace
Monitoring Commands
# Persistence status
INFO persistence
# Check last save
DEBUG SLEEP 0
LASTSAVE
# Force save and verify
BGSAVE
# Wait...
DEBUG RELOAD # Reload from disk (CAREFUL in prod)
# Check data integrity
DEBUG DIGEST
Assets
backup-redis.sh - Automated backup script
persistence-config.conf - Optimized config
References
PERSISTENCE_GUIDE.md - Strategy guide
Troubleshooting Guide
Common Issues & Solutions
1. BGSAVE Failing
Background saving error
Can't save in background: fork: Cannot allocate memory
# Disable THP
echo never > /sys/kernel/mm/transparent_hugepage/enabled
# Schedule BGSAVE during low traffic
save "" # Disable auto-save
# Use cron for scheduled saves
Debug Checklist
□ Persistence enabled? (INFO persistence)
□ Directory writable? (ls -la /var/lib/redis)
□ Disk space available? (df -h)
□ Memory for fork? (free -m)
□ Last save successful? (LASTSAVE)
□ AOF fsync configured? (CONFIG GET appendfsync)
□ RDB file valid? (redis-check-rdb)
□ AOF file valid? (redis-check-aof)
Performance Impact
Operation
Impact
Mitigation
BGSAVE
Memory spike (CoW)
Schedule low-traffic
AOF always
High latency
Use everysec
BGREWRITEAOF
CPU + I/O
Set min-size high
Recovery
Downtime
Use RDB preamble
Error Codes Reference
Code
Name
Description
Recovery
P001
FORK_FAIL
Cannot fork for BGSAVE
Enable overcommit
P002
RDB_CORRUPT
RDB file corrupted
Restore backup
P003
AOF_CORRUPT
AOF file corrupted
redis-check-aof --fix
P004
DISK_FULL
No space for persistence
Free disk space
P005
WRITE_ERR
Cannot write to disk
Check permissions
Test Template
# test_redis_persistence.pyimport redis
import pytest
import time
import subprocess
@pytest.fixturedefr():
return redis.Redis(decode_responses=True)
classTestRDB:
deftest_bgsave(self, r):
# Write some data
r.set("test:persist:1", "value1")
# Get last save time
last_save = r.lastsave()
# Trigger BGSAVE
r.bgsave()
# Wait for completion
time.sleep(2)
# Verify new save
new_save = r.lastsave()
assert new_save >= last_save
r.delete("test:persist:1")
deftest_persistence_info(self, r):
info = r.info("persistence")
assert"rdb_last_save_time"in info
assert"rdb_changes_since_last_save"in info
classTestAOF:
deftest_aof_rewrite(self, r):
info = r.info("persistence")
if info.get("aof_enabled"):
# Trigger rewrite
r.bgrewriteaof()
time.sleep(1)
# Check status
info = r.info("persistence")
assert info["aof_rewrite_in_progress"] in [0, 1]
classTestRecovery:
deftest_data_survives_restart(self, r):
"""Note: Requires actual Redis restart to fully test"""
key = "test:recovery:data"
r.set(key, "important_value")
r.bgsave()
time.sleep(1)
# In real test, would restart Redis here# For now, just verify data existsassert r.get(key) == "important_value"
r.delete(key)
