Redis Redis 4.0 Stable Release

eva2000 · Jul 15, 2017

Redis 4.0 Stable release is out now Redis. Centmin Mod installs Redis 3.2 branch from Remi YUM repo if you follow instructions outlined here. There is currently no Redis 4.0 YUM packages available in Remi YUM repo so would have to install Redis 4.0 yourself if you know how.

Release notes

https://raw.githubusercontent.com/antirez/redis/4.0/00-RELEASENOTES

================================================================================
Redis 4.0.0 Released Fri Jul 14 13:04:44 CEST 2017
================================================================================

Upgrade urgency CRITICAL: 4.0.0 GA fixes many important bugs.

Dear Redis users,

this is the first stable version of Redis 4.0. There are a number
of bug fixes and improvements compared to the previous RC, mainly:

* Different replication fixes to PSYNC2, the new 4.0 replication engine.
* Modules thread safe contexts were introduced. They are an experimental API right now, but the API is considered to be stable and usable when needed.
* SLOWLOG now logs the offending client name and address. Note that this is a backward compatibility breakage in case old code assumes that the slowlog entry is composed of exactly three entries.
* The modules native data types RDB format changed.
* The AOF check utility is now able to deal with RDB preambles.
* GEORADIUS_RO and GEORADIUSBYMEMBER_RO variants, not supporting the STORE option, were added in order to allow read-only scaling of such queries.
* HSET is now variadic, and HMSET is considered deprecated (but will be supported for years to come). Please use HSET in new code.
* GEORADIUS huge radius (>= ~6000 km) corner cases fixed, certain elements near the edges were not returned.
* DEBUG DIGEST modules API added.
* HyperLogLog commands no longer crash on certain input (non HLL) strings.
* Fixed SLAVEOF inside MULTI/EXEC blocks.
* Many other minor bug fixes and improvements.

Note that 4.0 is probably one of the most extreme releases of Redis ever
made in terms of changes inside the internals: all the aggregated data types
no longer use Redis Objects structures but directly SDS objects, certain
deletion operations are now threaded, the replication engine was modified
in many ways. So please handle this release with care. A few patch-level
releases will follow in the next weeks and months fixing the important issues
discovered by the users.

You can read the new set of features below in this file, there are a lot
of improvements that can make a real difference in real world use cases.
Also note that Redis 4.0 is, as usually, almost a perfect superset of Redis
3.2, so it is very rare that compatibility with the past is broken in terms
of exported commands.

IMPORTANT: Redis Cluster users, please note that, as specified in the list
of incompatibilities, Redis 4.0 cluster bus protocol is not compatible with
Redis 3.2, so in order to upgrade, a mass reboot of the instances is needed
and rolling upgrades are not possible. This change was needed in order to
add compatibility for Containers/NAT, where the bus port at a fixed offset
was not an acceptable design, so we had to change many things, resulting
in the incompatible protocol.

Have fun with Redis 4.0!
Salvatore
Click to expand...

From Google Groups

Dear Redis users,

finally after the canonical 1 year of developments, Redis 4.0.0 is out.
I'm very happy about 4.0, because this release finally overcomes
several very important limits of Redis. There are many things in the
4.2 plan, so Redis with 4.0 is not magically perfect. Yet it is a very
important, measurable, step forward, because of the following
features:

1) PSYNC2, the new replication engine. The way the replication
handshake and changes propagation happens between masters and slaves
was significantly changed. Now slaves promoted to masters are able to
accept the old slaves (reconfigured to point to the new master)
without the need to do a full resynchronization. Similarly slaves can
be stopped, upgraded and restarted, and can continue with the master
with just a partial resynchronization. People that run Redis
operations will appreciate this change...

2) Redis modules: finally it is possible to use Redis as a framework
to write networked services without reinventing everything from
scratch. Modules can extend Redis with new functionalities and data
structures, implement types which are exactly like the native types
(persisted in the RDB, rewritten in the AOF, ...), and there is
experimental support for blocking and threaded slow operations.

3) Redis is a better cache with 4.0, because it supports LFU other
than LRU as eviction algorithm, and in general the eviction algorithm
was improved in different ways.

4) DEL, FLUSHDB and FLUSHALL can now be ran in a different thread
calling UNLINK (instead of DEL), and using the ASYNC option of FLUSH
commands. This, together with the new SWAPDB command, open interesting
possibilities to atomically switch a given dataset to another one,
makes certain workloads using only DEL as O(N) commands a lot less
latency sensitive, and can improve also replication setup times,
because the slave, if configured to do so, can delete the old dataset
in background after a full resync.

5) It is now possible to enable a new AOF mode that rewrites the AOF
in terms of an RDB preamble + an AOF tail: faster rewrites, faster
reloads of the data, less latency issues.

6) The Raspberry Pi is now one a supported platform: every new Redis
release is tested on the Pi, the code was audited and fixed for
unaligned accesses, stack traces are supported on ARM.

7) The new MEMORY command, and especially MEMORY DOCTOR, allows to
have a better observability on the Redis memory usage. MEMORY USAGE
can precisely report a key memory utilization.

8) Redis Cluster was improved (but 4.2 will focus mostly on Cluster...
so the big news are headed): NAT/Containers support, better failure
detection (much lower traffic in big clusters), less CPU overhead in
cluster nodes compared to Redis 3.2.

9) Active memory defragmentation. Redis is able to defragment the
memory while online if the Jemalloc allocator is used (the default on
Linux). Useful for workloads where the allocator cannot keep the
fragmentation low enough, so the only possibility is for Redis and the
allocator to collaborate in order to defragment the memory.

A big thank you to the Redis community, providing help in issues, help
channels, fixing documentation, finding bugs, sending PRs. This Redis
release is surely the one that benefitted the most from external help.
A special thank you to my colleagues at Redis Labs that supported me
in all the possible ways, helped me in the hardest moments (see
Heisenbugs...), and that were extremely active in using, testing,
fixing and designing with me the modules API.

Please report 4.0.0 problems ASAP if you find any, I expect a few
patch releases coming before we actually reach the same level of
stability of Redis 3.2.

Enjoy!
Salvatore
Click to expand...

eva2000 · Jul 15, 2017

Redis 4.0 Custom Install Example

Did a quick custom source compiled Redis 4.0 install on CentOS 7.3 64bit with Redis 4.0 listening on custom TCP port 6479.

Code (Text):

ps aufxw | grep redis | grep -v grep
redis    14735  0.1  0.3 142892  5796 ?        Ssl  19:53   0:00 /usr/bin/redis-server 127.0.0.1:6379
redis    14808  0.1  0.1 145240  2628 ?        Ssl  19:55   0:00 /etc/redis6479/redis-server 127.0.0.1:6479

Source Installed Redis 4.0

Redis 4.0.0 listening on TCP port 6479

Code (Text):

/usr/local/bin/redis-server -v
Redis server v=4.0.0 sha=00000000:0 malloc=jemalloc-4.0.3 bits=64 build=c9d8b28825766954

Code (Text):

systemctl status redis6479
● redis6479.service - Redis persistent key-value database
   Loaded: loaded (/usr/lib/systemd/system/redis6479.service; enabled; vendor preset: disabled)
  Drop-In: /etc/systemd/system/redis6479.service.d
           └─limit.conf
   Active: active (running) since Fri 2017-07-14 19:55:40 UTC; 121ms ago
 Main PID: 14808 (redis-server)
   CGroup: /system.slice/redis6479.service
           └─14808 /etc/redis6479/redis-server 127.0.0.1:6479

Jul 14 19:55:40 centos7.localdomain systemd[1]: Started Redis persistent key-value database.
Jul 14 19:55:40 centos7.localdomain systemd[1]: Starting Redis persistent key-value database...

Code (Text):

redis-cli -h 127.0.0.1 -p 6479 info
# Server
redis_version:4.0.0
redis_git_sha1:00000000
redis_git_dirty:0
redis_build_id:c9d8b28825766954
redis_mode:standalone
os:Linux 3.10.0-514.26.2.el7.x86_64 x86_64
arch_bits:64
multiplexing_api:epoll
atomicvar_api:atomic-builtin
gcc_version:4.8.5
process_id:14808
run_id:f647eceb48306543f6b585d5a1364c3c2c5d8841
tcp_port:6479
uptime_in_seconds:268
uptime_in_days:0
hz:10
lru_clock:6890184
executable:/etc/redis6479/redis-server
config_file:/etc/redis6479/redis6479.conf

# Clients
connected_clients:1
client_longest_output_list:0
client_biggest_input_buf:0
blocked_clients:0

# Memory
used_memory:828536
used_memory_human:809.12K
used_memory_rss:2691072
used_memory_rss_human:2.57M
used_memory_peak:828536
used_memory_peak_human:809.12K
used_memory_peak_perc:108.21%
used_memory_overhead:815302
used_memory_startup:765672
used_memory_dataset:13234
used_memory_dataset_perc:21.05%
total_system_memory:1928650752
total_system_memory_human:1.80G
used_memory_lua:37888
used_memory_lua_human:37.00K
maxmemory:0
maxmemory_human:0B
maxmemory_policy:noeviction
mem_fragmentation_ratio:3.25
mem_allocator:jemalloc-4.0.3
active_defrag_running:0
lazyfree_pending_objects:0

# Persistence
loading:0
rdb_changes_since_last_save:0
rdb_bgsave_in_progress:0
rdb_last_save_time:1500062140
rdb_last_bgsave_status:ok
rdb_last_bgsave_time_sec:-1
rdb_current_bgsave_time_sec:-1
rdb_last_cow_size:0
aof_enabled:0
aof_rewrite_in_progress:0
aof_rewrite_scheduled:0
aof_last_rewrite_time_sec:-1
aof_current_rewrite_time_sec:-1
aof_last_bgrewrite_status:ok
aof_last_write_status:ok
aof_last_cow_size:0

# Stats
total_connections_received:1
total_commands_processed:0
instantaneous_ops_per_sec:0
total_net_input_bytes:14
total_net_output_bytes:0
instantaneous_input_kbps:0.00
instantaneous_output_kbps:0.00
rejected_connections:0
sync_full:0
sync_partial_ok:0
sync_partial_err:0
expired_keys:0
evicted_keys:0
keyspace_hits:0
keyspace_misses:0
pubsub_channels:0
pubsub_patterns:0
latest_fork_usec:0
migrate_cached_sockets:0
slave_expires_tracked_keys:0
active_defrag_hits:0
active_defrag_misses:0
active_defrag_key_hits:0
active_defrag_key_misses:0

# Replication
role:master
connected_slaves:0
master_replid:e4c672ccf795abecf9e14dddc16291fca490302e
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:0
second_repl_offset:-1
repl_backlog_active:0
repl_backlog_size:1048576
repl_backlog_first_byte_offset:0
repl_backlog_histlen:0

# CPU
used_cpu_sys:0.14
used_cpu_user:0.23
used_cpu_sys_children:0.00
used_cpu_user_children:0.00

# Cluster
cluster_enabled:0

# Keyspace

Remi YUM Redis 3.2.9

Redis 3.2.9 listening on TCP port 6379

Code (Text):

/usr/bin/redis-server -v  
Redis server v=3.2.9 sha=00000000:0 malloc=jemalloc-3.6.0 bits=64 build=a938017dd5f8fcc7

Code (Text):

systemctl status redis
● redis.service - Redis persistent key-value database
   Loaded: loaded (/usr/lib/systemd/system/redis.service; enabled; vendor preset: disabled)
  Drop-In: /etc/systemd/system/redis.service.d
           └─limit.conf
   Active: active (running) since Fri 2017-07-14 19:53:38 UTC; 3min 4s ago
 Main PID: 14735 (redis-server)
   CGroup: /system.slice/redis.service
           └─14735 /usr/bin/redis-server 127.0.0.1:6379

Jul 14 19:53:38 centos7.localdomain systemd[1]: Started Redis persistent key-value database.
Jul 14 19:53:38 centos7.localdomain systemd[1]: Starting Redis persistent key-value database...

Code (Text):

redis-cli info
# Server
redis_version:3.2.9
redis_git_sha1:00000000
redis_git_dirty:0
redis_build_id:a938017dd5f8fcc7
redis_mode:standalone
os:Linux 3.10.0-514.26.2.el7.x86_64 x86_64
arch_bits:64
multiplexing_api:epoll
gcc_version:4.8.5
process_id:14735
run_id:16336b79481362e66881ae649b605606e3d1d96d
tcp_port:6379
uptime_in_seconds:321
uptime_in_days:0
hz:10
lru_clock:6890115
executable:/usr/bin/redis-server
config_file:/etc/redis.conf

# Clients
connected_clients:1
client_longest_output_list:0
client_biggest_input_buf:0
blocked_clients:0

# Memory
used_memory:813224
used_memory_human:794.16K
used_memory_rss:5935104
used_memory_rss_human:5.66M
used_memory_peak:813224
used_memory_peak_human:794.16K
total_system_memory:1928650752
total_system_memory_human:1.80G
used_memory_lua:37888
used_memory_lua_human:37.00K
maxmemory:0
maxmemory_human:0B
maxmemory_policy:noeviction
mem_fragmentation_ratio:7.30
mem_allocator:jemalloc-3.6.0

# Persistence
loading:0
rdb_changes_since_last_save:0
rdb_bgsave_in_progress:0
rdb_last_save_time:1500062018
rdb_last_bgsave_status:ok
rdb_last_bgsave_time_sec:-1
rdb_current_bgsave_time_sec:-1
aof_enabled:0
aof_rewrite_in_progress:0
aof_rewrite_scheduled:0
aof_last_rewrite_time_sec:-1
aof_current_rewrite_time_sec:-1
aof_last_bgrewrite_status:ok
aof_last_write_status:ok

# Stats
total_connections_received:1
total_commands_processed:0
instantaneous_ops_per_sec:0
total_net_input_bytes:14
total_net_output_bytes:0
instantaneous_input_kbps:0.00
instantaneous_output_kbps:0.00
rejected_connections:0
sync_full:0
sync_partial_ok:0
sync_partial_err:0
expired_keys:0
evicted_keys:0
keyspace_hits:0
keyspace_misses:0
pubsub_channels:0
pubsub_patterns:0
latest_fork_usec:0
migrate_cached_sockets:0

# Replication
role:master
connected_slaves:0
master_repl_offset:0
repl_backlog_active:0
repl_backlog_size:1048576
repl_backlog_first_byte_offset:0
repl_backlog_histlen:0

# CPU
used_cpu_sys:0.19
used_cpu_user:0.32
used_cpu_sys_children:0.00
used_cpu_user_children:0.00

# Cluster
cluster_enabled:0

# Keyspace

Redis Benchmarks

Compare standalone Redis 3.2.9 vs Redis 4.0 standalone

redis version	set	get	mset
redis 3.2.9	6060.61	4901.96	1872.66
redis 4.0.0	5747.13	5076.14	2109.70

Code (Text):

PORTS='6379 6479'
for PORT in $PORTS; do echo "redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t set -P 1000 -c 100"; redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t set -P 1000 -c 100; done

Code (Text):

redis-benchmark -h 127.0.0.1 -p 6379 -n 1000 -r 1000 -t set -P 1000 -c 100
====== SET ======
  1000 requests completed in 0.17 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 12 milliseconds
36.50% <= 19 milliseconds
72.90% <= 20 milliseconds
100.00% <= 20 milliseconds
6060.61 requests per second


redis-benchmark -h 127.0.0.1 -p 6479 -n 1000 -r 1000 -t set -P 1000 -c 100
====== SET ======
  1000 requests completed in 0.17 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 14 milliseconds
36.50% <= 24 milliseconds
100.00% <= 24 milliseconds
5747.13 requests per second

Code (Text):

PORTS='6379 6479'
for PORT in $PORTS; do echo "redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t get -P 1000 -c 100"; redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t get -P 1000 -c 100; done

Code (Text):

redis-benchmark -h 127.0.0.1 -p 6379 -n 1000 -r 1000 -t get -P 1000 -c 100
====== GET ======
  1000 requests completed in 0.20 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 10 milliseconds
45.60% <= 40 milliseconds
100.00% <= 40 milliseconds
4901.96 requests per second


redis-benchmark -h 127.0.0.1 -p 6479 -n 1000 -r 1000 -t get -P 1000 -c 100
====== GET ======
  1000 requests completed in 0.20 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 12 milliseconds
45.60% <= 35 milliseconds
100.00% <= 35 milliseconds
5076.14 requests per second

Code (Text):

PORTS='6379 6479'
for PORT in $PORTS; do echo "redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t mset -P 1000 -c 100"; redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t mset -P 1000 -c 100; done

Code (Text):

redis-benchmark -h 127.0.0.1 -p 6379 -n 1000 -r 1000 -t mset -P 1000 -c 100
====== MSET (10 keys) ======
  1000 requests completed in 0.53 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 66 milliseconds
21.90% <= 67 milliseconds
56.10% <= 68 milliseconds
100.00% <= 68 milliseconds
1872.66 requests per second


redis-benchmark -h 127.0.0.1 -p 6479 -n 1000 -r 1000 -t mset -P 1000 -c 100
====== MSET (10 keys) ======
  1000 requests completed in 0.47 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 65 milliseconds
12.10% <= 66 milliseconds
51.20% <= 67 milliseconds
100.00% <= 67 milliseconds
2109.70 requests per second

eva2000 · Jul 15, 2017

Redis 4.0 Master / Slave Replication

Setup local Redis 4.0 master/slave replication on TCP ports 6479 (master) and 6480 (slave) + 3x redis sentinel with quorum of 2) tied to master (6479)

Configured as follows:

redis systemd files at /usr/lib/systemd/system/redis6479.service and /usr/lib/systemd/system/redis6480.service
redis-server binaries at /etc/redis6479/redis-server and /etc/redis6480/redis-server copied from /usr/local/bin/redis-server (source install) instead of /usr/bin/redis-server
redis pid files at /var/run/redis/redis_6479.pid and /var/run/redis/redis_6480.pid
dedicated redis config files at /etc/redis6479/redis6479.conf and /etc/redis6479/redis6480.conf
each config file will have dbfilename as dump6479.rdb and dump6480.rdb with commented out unix sockets at /var/run/redis/redis6479.sock and /var/run/redis/redis6480.sock
dedicated redis data directories at /var/lib/redis6479 and /var/lib/redis6480
dedicated redis log files at /var/log/redis/redis6479.log and /var/log/redis/redis6480.log
sentinel config file /root/tools/redis-sentinel/sentinel-16479.conf where STARTPORT is 6479 and incremented by 2 for /root/tools/redis-sentinel/sentinel-16480.conf and /root/tools/redis-sentinel/sentinel-16481.conf
sentinel port is STARTPORT + 10000 = 6479 + 10000 = 16479 and incremented by 2 for 16480 and 16481
sentinel init.d startup script is at /etc/init.d/sentinel_16479 and incremented by 2 for /etc/init.d/sentinel_16480 and /etc/init.d/sentinel_16481
sentinel directory /var/lib/redis/sentinel_16479, /var/lib/redis/sentinel_16480 and /var/lib/redis/sentinel_16481
sentinel log /var/log/redis/sentinel-16479.log, /var/log/redis/sentinel-16480.log and /var/log/redis/sentinel-16481.log
sentinel pid file /var/run/redis/redis-sentinel-16479.pid, /var/run/redis/redis-sentinel-16480.pid and /var/run/redis/redis-sentinel-16481.pid
quorum is set to 2 with 3x redis sentinels

Code (Text):

/etc/redis6479/redis-server -v
Redis server v=4.0.0 sha=00000000:0 malloc=jemalloc-4.0.3 bits=64 build=269fa3414af1c24f

/etc/redis6480/redis-server -v
Redis server v=4.0.0 sha=00000000:0 malloc=jemalloc-4.0.3 bits=64 build=269fa3414af1c24f

Code (Text):

ps aufxw | grep redis | grep -v grep
redis    10794  0.2  0.3 142892  5796 ?        Ssl  20:14   0:01 /usr/bin/redis-server 127.0.0.1:6379
redis    15557  0.2  0.1 147288  2816 ?        Ssl  20:25   0:00 /etc/redis6479/redis-server 127.0.0.1:6479
root     15625  0.5  0.1  41652  2340 ?        Ssl  20:25   0:00 /etc/redis6479/redis-server *:16479 [sentinel]
root     15673  0.5  0.1  41652  2328 ?        Ssl  20:25   0:00 /etc/redis6479/redis-server *:16480 [sentinel]
root     15721  0.3  0.1  41652  2268 ?        Ssl  20:25   0:00 /etc/redis6479/redis-server *:16481 [sentinel]
redis    15773  0.5  0.1 147288  2696 ?        Ssl  20:25   0:00 /etc/redis6480/redis-server 127.0.0.1:6480

Code (Text):

/usr/local/bin/redis-cli -h 127.0.0.1 -p 6479 INFO
# Server
redis_version:4.0.0
redis_git_sha1:00000000
redis_git_dirty:0
redis_build_id:269fa3414af1c24f
redis_mode:standalone
os:Linux 3.10.0-514.26.2.el7.x86_64 x86_64
arch_bits:64
multiplexing_api:epoll
atomicvar_api:atomic-builtin
gcc_version:4.8.5
process_id:15557
run_id:290bedce2c783b9d7eec67e12a78261ea9f78236
tcp_port:6479
uptime_in_seconds:114
uptime_in_days:0
hz:10
lru_clock:6891826
executable:/etc/redis6479/redis-server
config_file:/etc/redis6479/redis6479.conf

# Clients
connected_clients:7
client_longest_output_list:0
client_biggest_input_buf:0
blocked_clients:0

# Memory
used_memory:2023000
used_memory_human:1.93M
used_memory_rss:2883584
used_memory_rss_human:2.75M
used_memory_peak:2123912
used_memory_peak_human:2.03M
used_memory_peak_perc:95.25%
used_memory_overhead:1981884
used_memory_startup:765672
used_memory_dataset:41116
used_memory_dataset_perc:3.27%
total_system_memory:1928650752
total_system_memory_human:1.80G
used_memory_lua:37888
used_memory_lua_human:37.00K
maxmemory:0
maxmemory_human:0B
maxmemory_policy:noeviction
mem_fragmentation_ratio:1.43
mem_allocator:jemalloc-4.0.3
active_defrag_running:0
lazyfree_pending_objects:0

# Persistence
loading:0
rdb_changes_since_last_save:0
rdb_bgsave_in_progress:0
rdb_last_save_time:1500063945
rdb_last_bgsave_status:ok
rdb_last_bgsave_time_sec:0
rdb_current_bgsave_time_sec:-1
rdb_last_cow_size:217088
aof_enabled:0
aof_rewrite_in_progress:0
aof_rewrite_scheduled:0
aof_last_rewrite_time_sec:-1
aof_current_rewrite_time_sec:-1
aof_last_bgrewrite_status:ok
aof_last_write_status:ok
aof_last_cow_size:0

# Stats
total_connections_received:10
total_commands_processed:630
instantaneous_ops_per_sec:7
total_net_input_bytes:31081
total_net_output_bytes:183508
instantaneous_input_kbps:0.35
instantaneous_output_kbps:2.72
rejected_connections:0
sync_full:1
sync_partial_ok:0
sync_partial_err:0
expired_keys:0
evicted_keys:0
keyspace_hits:0
keyspace_misses:0
pubsub_channels:1
pubsub_patterns:0
latest_fork_usec:135
migrate_cached_sockets:0
slave_expires_tracked_keys:0
active_defrag_hits:0
active_defrag_misses:0
active_defrag_key_hits:0
active_defrag_key_misses:0

# Replication
role:master
connected_slaves:1
slave0:ip=127.0.0.1,port=6480,state=online,offset=20985,lag=0
master_replid:62110e54ecc48592cc27d6474ffa913d40b503cc
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:20985
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:20985

# CPU
used_cpu_sys:0.18
used_cpu_user:0.26
used_cpu_sys_children:0.00
used_cpu_user_children:0.00

# Cluster
cluster_enabled:0

# Keyspace

Code (Text):

/usr/local/bin/redis-cli -h 127.0.0.1 -p 6480 INFO
# Server
redis_version:4.0.0
redis_git_sha1:00000000
redis_git_dirty:0
redis_build_id:269fa3414af1c24f
redis_mode:standalone
os:Linux 3.10.0-514.26.2.el7.x86_64 x86_64
arch_bits:64
multiplexing_api:epoll
atomicvar_api:atomic-builtin
gcc_version:4.8.5
process_id:15773
run_id:f8ac7fdf44a076e31d91229a1cf243e8f00f8aa6
tcp_port:6480
uptime_in_seconds:132
uptime_in_days:0
hz:10
lru_clock:6891853
executable:/etc/redis6480/redis-server
config_file:/etc/redis6480/redis6480.conf

# Clients
connected_clients:8
client_longest_output_list:0
client_biggest_input_buf:0
blocked_clients:0

# Memory
used_memory:2023336
used_memory_human:1.93M
used_memory_rss:2760704
used_memory_rss_human:2.63M
used_memory_peak:2124296
used_memory_peak_human:2.03M
used_memory_peak_perc:95.25%
used_memory_overhead:1981900
used_memory_startup:765688
used_memory_dataset:41436
used_memory_dataset_perc:3.29%
total_system_memory:1928650752
total_system_memory_human:1.80G
used_memory_lua:37888
used_memory_lua_human:37.00K
maxmemory:0
maxmemory_human:0B
maxmemory_policy:noeviction
mem_fragmentation_ratio:1.36
mem_allocator:jemalloc-4.0.3
active_defrag_running:0
lazyfree_pending_objects:0

# Persistence
loading:0
rdb_changes_since_last_save:0
rdb_bgsave_in_progress:0
rdb_last_save_time:1500063945
rdb_last_bgsave_status:ok
rdb_last_bgsave_time_sec:-1
rdb_current_bgsave_time_sec:-1
rdb_last_cow_size:0
aof_enabled:0
aof_rewrite_in_progress:0
aof_rewrite_scheduled:0
aof_last_rewrite_time_sec:-1
aof_current_rewrite_time_sec:-1
aof_last_bgrewrite_status:ok
aof_last_write_status:ok
aof_last_cow_size:0

# Stats
total_connections_received:9
total_commands_processed:806
instantaneous_ops_per_sec:6
total_net_input_bytes:57619
total_net_output_bytes:271717
instantaneous_input_kbps:0.54
instantaneous_output_kbps:1.53
rejected_connections:0
sync_full:0
sync_partial_ok:0
sync_partial_err:0
expired_keys:0
evicted_keys:0
keyspace_hits:0
keyspace_misses:0
pubsub_channels:1
pubsub_patterns:0
latest_fork_usec:0
migrate_cached_sockets:0
slave_expires_tracked_keys:0
active_defrag_hits:0
active_defrag_misses:0
active_defrag_key_hits:0
active_defrag_key_misses:0

# Replication
role:slave
master_host:127.0.0.1
master_port:6479
master_link_status:up
master_last_io_seconds_ago:0
master_sync_in_progress:0
slave_repl_offset:26317
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:62110e54ecc48592cc27d6474ffa913d40b503cc
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:26317
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:26317

# CPU
used_cpu_sys:0.21
used_cpu_user:0.34
used_cpu_sys_children:0.00
used_cpu_user_children:0.00

# Cluster
cluster_enabled:0

# Keyspace

Redis 4.0 replication master log

Code (Text):

tail -1000 /var/log/redis/redis6479.log
15557:C 14 Jul 20:25:36.784 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
15557:C 14 Jul 20:25:36.784 # Redis version=4.0.0, bits=64, commit=00000000, modified=0, pid=15557, just started
15557:C 14 Jul 20:25:36.784 # Configuration loaded
15557:M 14 Jul 20:25:36.787 * Running mode=standalone, port=6479.
15557:M 14 Jul 20:25:36.787 # Server initialized
15557:M 14 Jul 20:25:36.787 * Ready to accept connections
15557:M 14 Jul 20:25:45.503 * Slave 127.0.0.1:6480 asks for synchronization
15557:M 14 Jul 20:25:45.503 * Full resync requested by slave 127.0.0.1:6480
15557:M 14 Jul 20:25:45.503 * Starting BGSAVE for SYNC with target: disk
15557:M 14 Jul 20:25:45.503 * Background saving started by pid 15777
15777:C 14 Jul 20:25:45.526 * DB saved on disk
15777:C 14 Jul 20:25:45.526 * RDB: 0 MB of memory used by copy-on-write
15557:M 14 Jul 20:25:45.557 * Background saving terminated with success
15557:M 14 Jul 20:25:45.558 * Synchronization with slave 127.0.0.1:6480 succeeded

Redis 4.0 replication slave log

Code (Text):

tail -1000 /var/log/redis/redis6480.log
15773:C 14 Jul 20:25:45.499 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
15773:C 14 Jul 20:25:45.500 # Redis version=4.0.0, bits=64, commit=00000000, modified=0, pid=15773, just started
15773:C 14 Jul 20:25:45.500 # Configuration loaded
15773:S 14 Jul 20:25:45.502 * Running mode=standalone, port=6480.
15773:S 14 Jul 20:25:45.502 # Server initialized
15773:S 14 Jul 20:25:45.502 * Ready to accept connections
15773:S 14 Jul 20:25:45.502 * Connecting to MASTER 127.0.0.1:6479
15773:S 14 Jul 20:25:45.502 * MASTER <-> SLAVE sync started
15773:S 14 Jul 20:25:45.502 * Non blocking connect for SYNC fired the event.
15773:S 14 Jul 20:25:45.502 * Master replied to PING, replication can continue...
15773:S 14 Jul 20:25:45.503 * Partial resynchronization not possible (no cached master)
15773:S 14 Jul 20:25:45.503 * Full resync from master: 62110e54ecc48592cc27d6474ffa913d40b503cc:0
15773:S 14 Jul 20:25:45.558 * MASTER <-> SLAVE sync: receiving 175 bytes from master
15773:S 14 Jul 20:25:45.558 * MASTER <-> SLAVE sync: Flushing old data
15773:S 14 Jul 20:25:45.558 * MASTER <-> SLAVE sync: Loading DB in memory
15773:S 14 Jul 20:25:45.558 * MASTER <-> SLAVE sync: Finished with success

Redis 4.0 replication sentinel log for TCP 16479

Code (Text):

tail -1000 /var/log/redis/sentinel-16479.log
15620:X 14 Jul 20:25:39.044 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
15620:X 14 Jul 20:25:39.044 # Redis version=4.0.0, bits=64, commit=00000000, modified=0, pid=15620, just started
15620:X 14 Jul 20:25:39.044 # Configuration loaded
15625:X 14 Jul 20:25:39.046 * Running mode=sentinel, port=16479.
15625:X 14 Jul 20:25:39.059 # Sentinel ID is 10e60db015c9750647560096bcb7e94e634b7ccd
15625:X 14 Jul 20:25:39.059 # +monitor master master-6479 127.0.0.1 6479 quorum 2
15625:X 14 Jul 20:25:43.223 * +sentinel sentinel 88d8236c5dc5eace0f92458ad828be4e63aae715 127.0.0.1 16480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:25:45.382 * +sentinel sentinel 292bf58b986c4c36699eb35efd1de36c8c7ba307 127.0.0.1 16481 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:25:49.140 * +slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479

Simulate Redis master on port 6479 failure

Code (Text):

kill -9 $(cat /var/run/redis/redis_6479.pid)

Code (Text):

redis-cli -h 127.0.0.1 -p 6479 INFO REPLICATION
Could not connect to Redis at 127.0.0.1:6479: Connection refused

Check replication info for promoted redis slave 6480 to master where role = master

Code (Text):

redis-cli -h 127.0.0.1 -p 6480 INFO REPLICATION
# Replication
role:master
connected_slaves:0
master_replid:ec956cf16814ac18d3e705bb3adece5fe04fe4e4
master_replid2:62110e54ecc48592cc27d6474ffa913d40b503cc
master_repl_offset:102446
second_repl_offset:95760
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:102446

Checking sentinel #1 log on port 16479

Code (Text):

tail -50 /var/log/redis/sentinel-16479.log
15620:X 14 Jul 20:25:39.044 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
15620:X 14 Jul 20:25:39.044 # Redis version=4.0.0, bits=64, commit=00000000, modified=0, pid=15620, just started
15620:X 14 Jul 20:25:39.044 # Configuration loaded
15625:X 14 Jul 20:25:39.046 * Running mode=sentinel, port=16479.
15625:X 14 Jul 20:25:39.059 # Sentinel ID is 10e60db015c9750647560096bcb7e94e634b7ccd
15625:X 14 Jul 20:25:39.059 # +monitor master master-6479 127.0.0.1 6479 quorum 2
15625:X 14 Jul 20:25:43.223 * +sentinel sentinel 88d8236c5dc5eace0f92458ad828be4e63aae715 127.0.0.1 16480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:25:45.382 * +sentinel sentinel 292bf58b986c4c36699eb35efd1de36c8c7ba307 127.0.0.1 16481 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:25:49.140 * +slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.580 # +sdown master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.663 # +odown master master-6479 127.0.0.1 6479 #quorum 2/2
15625:X 14 Jul 20:33:49.664 # +new-epoch 1
15625:X 14 Jul 20:33:49.664 # +try-failover master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.667 # +vote-for-leader 10e60db015c9750647560096bcb7e94e634b7ccd 1
15625:X 14 Jul 20:33:49.675 # 292bf58b986c4c36699eb35efd1de36c8c7ba307 voted for 10e60db015c9750647560096bcb7e94e634b7ccd 1
15625:X 14 Jul 20:33:49.675 # 88d8236c5dc5eace0f92458ad828be4e63aae715 voted for 10e60db015c9750647560096bcb7e94e634b7ccd 1
15625:X 14 Jul 20:33:49.730 # +elected-leader master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.730 # +failover-state-select-slave master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.822 # +selected-slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.822 * +failover-state-send-slaveof-noone slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.888 * +failover-state-wait-promotion slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.691 # +promoted-slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.691 # +failover-state-reconf-slaves master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.790 # +failover-end master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.791 # +switch-master master-6479 127.0.0.1 6479 127.0.0.1 6480
15625:X 14 Jul 20:33:50.791 * +slave slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480
15625:X 14 Jul 20:33:53.808 # +sdown slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480

Checking sentinel #2 log on port 16480

Code (Text):

tail -50 /var/log/redis/sentinel-16480.log
15668:X 14 Jul 20:25:41.168 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
15668:X 14 Jul 20:25:41.168 # Redis version=4.0.0, bits=64, commit=00000000, modified=0, pid=15668, just started
15668:X 14 Jul 20:25:41.168 # Configuration loaded
15673:X 14 Jul 20:25:41.171 * Running mode=sentinel, port=16480.
15673:X 14 Jul 20:25:41.174 # Sentinel ID is 88d8236c5dc5eace0f92458ad828be4e63aae715
15673:X 14 Jul 20:25:41.174 # +monitor master master-6479 127.0.0.1 6479 quorum 2
15673:X 14 Jul 20:25:43.156 * +sentinel sentinel 10e60db015c9750647560096bcb7e94e634b7ccd 127.0.0.1 16479 @ master-6479 127.0.0.1 6479
15673:X 14 Jul 20:25:45.382 * +sentinel sentinel 292bf58b986c4c36699eb35efd1de36c8c7ba307 127.0.0.1 16481 @ master-6479 127.0.0.1 6479
15673:X 14 Jul 20:25:51.189 * +slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15673:X 14 Jul 20:33:49.616 # +sdown master master-6479 127.0.0.1 6479
15673:X 14 Jul 20:33:49.672 # +new-epoch 1
15673:X 14 Jul 20:33:49.674 # +vote-for-leader 10e60db015c9750647560096bcb7e94e634b7ccd 1
15673:X 14 Jul 20:33:49.675 # +odown master master-6479 127.0.0.1 6479 #quorum 3/2
15673:X 14 Jul 20:33:49.675 # Next failover delay: I will not start a failover before Fri Jul 14 20:34:01 2017
15673:X 14 Jul 20:33:50.793 # +config-update-from sentinel 10e60db015c9750647560096bcb7e94e634b7ccd 127.0.0.1 16479 @ master-6479 127.0.0.1 6479
15673:X 14 Jul 20:33:50.793 # +switch-master master-6479 127.0.0.1 6479 127.0.0.1 6480
15673:X 14 Jul 20:33:50.793 * +slave slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480
15673:X 14 Jul 20:33:53.834 # +sdown slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480

Checking sentinel #3 log on port 16481

Code (Text):

tail -50 /var/log/redis/sentinel-16481.log
15716:X 14 Jul 20:25:43.301 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
15716:X 14 Jul 20:25:43.302 # Redis version=4.0.0, bits=64, commit=00000000, modified=0, pid=15716, just started
15716:X 14 Jul 20:25:43.302 # Configuration loaded
15721:X 14 Jul 20:25:43.303 * Running mode=sentinel, port=16481.
15721:X 14 Jul 20:25:43.309 # Sentinel ID is 292bf58b986c4c36699eb35efd1de36c8c7ba307
15721:X 14 Jul 20:25:43.309 # +monitor master master-6479 127.0.0.1 6479 quorum 2
15721:X 14 Jul 20:25:45.202 * +sentinel sentinel 10e60db015c9750647560096bcb7e94e634b7ccd 127.0.0.1 16479 @ master-6479 127.0.0.1 6479
15721:X 14 Jul 20:25:45.298 * +sentinel sentinel 88d8236c5dc5eace0f92458ad828be4e63aae715 127.0.0.1 16480 @ master-6479 127.0.0.1 6479
15721:X 14 Jul 20:25:53.370 * +slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15721:X 14 Jul 20:33:49.553 # +sdown master master-6479 127.0.0.1 6479
15721:X 14 Jul 20:33:49.672 # +new-epoch 1
15721:X 14 Jul 20:33:49.674 # +vote-for-leader 10e60db015c9750647560096bcb7e94e634b7ccd 1
15721:X 14 Jul 20:33:50.653 # +odown master master-6479 127.0.0.1 6479 #quorum 3/2
15721:X 14 Jul 20:33:50.654 # Next failover delay: I will not start a failover before Fri Jul 14 20:34:01 2017
15721:X 14 Jul 20:33:50.793 # +config-update-from sentinel 10e60db015c9750647560096bcb7e94e634b7ccd 127.0.0.1 16479 @ master-6479 127.0.0.1 6479
15721:X 14 Jul 20:33:50.793 # +switch-master master-6479 127.0.0.1 6479 127.0.0.1 6480
15721:X 14 Jul 20:33:50.793 * +slave slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480
15721:X 14 Jul 20:33:53.815 # +sdown slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480

Restart Redis 6479 port instance

Code (Text):

systemctl start redis6479

Redis 6479 role = slave

Code (Text):

redis-cli -h 127.0.0.1 -p 6479 INFO REPLICATION
# Replication
role:slave
master_host:127.0.0.1
master_port:6480
master_link_status:up
master_last_io_seconds_ago:0
master_sync_in_progress:0
slave_repl_offset:171758
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:ec956cf16814ac18d3e705bb3adece5fe04fe4e4
master_replid2:62110e54ecc48592cc27d6474ffa913d40b503cc
master_repl_offset:0
second_repl_offset:1
repl_backlog_active:0
repl_backlog_size:1048576
repl_backlog_first_byte_offset:0
repl_backlog_histlen:0

Redis 6480 role = master

Code (Text):

redis-cli -h 127.0.0.1 -p 6480 INFO REPLICATION
# Replication
role:master
connected_slaves:1
slave0:ip=127.0.0.1,port=6479,state=online,offset=178886,lag=0
master_replid:ec956cf16814ac18d3e705bb3adece5fe04fe4e4
master_replid2:62110e54ecc48592cc27d6474ffa913d40b503cc
master_repl_offset:178886
second_repl_offset:95760
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:178886

Sentinel log at /var/log/redis/sentinel-16479.log

Code (Text):

tail -10 /var/log/redis/sentinel-16479.log    
15625:X 14 Jul 20:33:49.822 # +selected-slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.822 * +failover-state-send-slaveof-noone slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:49.888 * +failover-state-wait-promotion slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.691 # +promoted-slave slave 127.0.0.1:6480 127.0.0.1 6480 @ master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.691 # +failover-state-reconf-slaves master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.790 # +failover-end master master-6479 127.0.0.1 6479
15625:X 14 Jul 20:33:50.791 # +switch-master master-6479 127.0.0.1 6479 127.0.0.1 6480
15625:X 14 Jul 20:33:50.791 * +slave slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480
15625:X 14 Jul 20:33:53.808 # +sdown slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480
15625:X 14 Jul 20:37:16.051 # -sdown slave 127.0.0.1:6479 127.0.0.1 6479 @ master-6479 127.0.0.1 6480

Redis Benchmarks

Compare standalone Redis 3.2.9 vs Redis 4.0 replication master/slave. Replication usually would be slower due to writes to slave.

redis version	set	get	mset
redis 3.2.9	6289.31	5208.33	1697.79
redis 4.0.0	3968.25	4484.30	873.36

Code (Text):

PORTS='6379 6479'
for PORT in $PORTS; do echo "redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t set -P 1000 -c 100"; redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t set -P 1000 -c 100; done

Code (Text):

redis-benchmark -h 127.0.0.1 -p 6379 -n 1000 -r 1000 -t set -P 1000 -c 100
====== SET ======
  1000 requests completed in 0.16 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 12 milliseconds
36.50% <= 19 milliseconds
72.90% <= 20 milliseconds
100.00% <= 20 milliseconds
6289.31 requests per second


redis-benchmark -h 127.0.0.1 -p 6479 -n 1000 -r 1000 -t set -P 1000 -c 100
====== SET ======
  1000 requests completed in 0.25 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 20 milliseconds
36.50% <= 43 milliseconds
100.00% <= 43 milliseconds
3968.25 requests per second

Code (Text):

PORTS='6379 6479'
for PORT in $PORTS; do echo "redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t get -P 1000 -c 100"; redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t get -P 1000 -c 100; done

Code (Text):

redis-benchmark -h 127.0.0.1 -p 6379 -n 1000 -r 1000 -t get -P 1000 -c 100
====== GET ======
  1000 requests completed in 0.19 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 10 milliseconds
45.60% <= 34 milliseconds
100.00% <= 34 milliseconds
5208.33 requests per second


redis-benchmark -h 127.0.0.1 -p 6479 -n 1000 -r 1000 -t get -P 1000 -c 100
====== GET ======
  1000 requests completed in 0.22 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 10 milliseconds
45.60% <= 41 milliseconds
100.00% <= 41 milliseconds
4484.30 requests per second

Code (Text):

PORTS='6379 6479'
for PORT in $PORTS; do echo "redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t mset -P 1000 -c 100"; redis-benchmark -h 127.0.0.1 -p $PORT -n 1000 -r 1000 -t mset -P 1000 -c 100; done

Code (Text):

redis-benchmark -h 127.0.0.1 -p 6379 -n 1000 -r 1000 -t mset -P 1000 -c 100
====== MSET (10 keys) ======
  1000 requests completed in 0.59 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 95 milliseconds
21.90% <= 97 milliseconds
100.00% <= 97 milliseconds
1697.79 requests per second


redis-benchmark -h 127.0.0.1 -p 6479 -n 1000 -r 1000 -t mset -P 1000 -c 100
====== MSET (10 keys) ======
  1000 requests completed in 1.14 seconds
  100 parallel clients
  3 bytes payload
  keep alive: 1

0.10% <= 66 milliseconds
7.40% <= 69 milliseconds
26.90% <= 70 milliseconds
46.40% <= 71 milliseconds
100.00% <= 71 milliseconds
873.36 requests per second

eva2000 · Jul 15, 2017

Redis 4.0 Replication PSYNC2

From The first release candidate of Redis 4.0 is out - <antirez>

2. Replication version 2

So, that’s going to be very useful in production from the POV of operations. At some point in the past we introduced what is known as “PSYNC”. It was a new master-slave protocol that allowed the master and the salve to continue from where they were, if the connection between the two broke. Before that, every single connection break in the replication link between master and slave would result into a full synchronization: generate an RDB file in the master, transfer it, load it in the slave, ok you know how this works. So PSYNC was like a real improvement. But not enough…

PSYNC was not good enough when there was a failover. If a slave is promoted to master, the slaves that replicated with the old master were not able to connect to the new promoted slave and PSYNC with it: a full resynchronization was needed. This is not great, and is not good for Redis Cluster as well. However fixing this required to do changes to the replication protocol, because I really wanted to make sure that partial resyncs where going to work after any possible topology change, as long as there was a common replication history among the instances.

So the first change needed was about how “chained replication” works, that is, slaves of slaves of slaves … How do they work? Like, A is the master, and we have something like that:

A —> B —> C —> D

So A is the master of B but B is the master of C and so forth. Before Redis 4.0 what was happening is that B received the replication protocol from A. The replication protocol is, normally, a stream of write commands. B acted as a master for C just doing, internally, what A was doing: at every write it was generating again a suitable replication protocol to pass to C, and so forth.

Now instead B just proxies, verbatim, what it receives from A to C, and C will do the same for D: given that all the sub slaves now receive an identical stream of bytes, they can “tag” a given history, and use the tag and the offset to always try to continue if they have something in common.

The master itself, when it is turned into a slave, is now able to PSYNC with the new master. And slaves can usually PSYNC with the master even after a “clean” restart, using the info inside the RDB file, that now stores the replication tags and offsets.

Details are more complex than that, in order to make it working well, but well the deal here is, don’t be annoyed by full resynchronizations if possible. And PSYNC v2 does this well apparently. Please try it and let me know if you are interested in this feature.
Click to expand...

Log in / Register

Forums

Join the community today

Redis Redis 4.0 Stable Release

eva2000 Administrator Staff Member

eva2000 Administrator Staff Member

Redis 4.0 Custom Install Example

Redis Benchmarks

eva2000 Administrator Staff Member

Redis 4.0 Master / Slave Replication

Redis Benchmarks

eva2000 Administrator Staff Member

Redis 4.0 Replication PSYNC2

.

Log in / Register

Useful Searches

Join the community today

Redis Redis 4.0 Stable Release

eva2000 Administrator Staff Member

eva2000 Administrator Staff Member

Redis 4.0 Custom Install Example

Redis Benchmarks

eva2000 Administrator Staff Member

Redis 4.0 Master / Slave Replication

Redis Benchmarks

eva2000 Administrator Staff Member

Redis 4.0 Replication PSYNC2

.