Skip to content

Hitless upgrades #3142

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 15 commits into
base: master
Choose a base branch
from
+971 −109
Draft

Hitless upgrades #3142

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
5d80d07
handle smigrating
nkaradzhov Nov 11, 2025
4ec6771
first approximation to handling smigrated
nkaradzhov Nov 13, 2025
e186a13
deduplicate notifications based on sequence id
nkaradzhov Dec 2, 2025
7648abb
add slotnumber to commands
nkaradzhov Dec 2, 2025
8abb22d
add support for extracting commands from queue
nkaradzhov Dec 1, 2025
7b3a1fd
parse notification
nkaradzhov Dec 1, 2025
9b8177e
work on main algo
nkaradzhov Dec 2, 2025
b3fb5db
fix: handle string values in push message reply comparison
nkaradzhov Dec 2, 2025
a8a4525
parse SMIGRATED according to new format
nkaradzhov Dec 2, 2025
dbe604a
comply with the new notification structure
nkaradzhov Dec 2, 2025
6da583a
refine algo
nkaradzhov Dec 2, 2025
c70dbb4
handle pubSubNode replacement
nkaradzhov Dec 3, 2025
a265d79
tests: merge all `after` functions into one
nkaradzhov Dec 5, 2025
2f39426
tests: add `testWithProxiedCluster()` function
nkaradzhov Dec 5, 2025
9e052d3
tests: add ProxyController for easier proxy comms
nkaradzhov Dec 5, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
175 changes: 175 additions & 0 deletions CLUSTER-ARCHITECTURE.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,175 @@
Redis Cluster Architecture Diagram
===================================

┌─────────────────────────────────────────┐
│ RedisCluster (Root) │
│ │
│ - _options: RedisClusterOptions │
│ - _slots: RedisClusterSlots │
│ - _commandOptions │
│ │
│ Methods: │
│ + connect() │
│ + sendCommand() │
│ + MULTI() │
│ + SUBSCRIBE() / UNSUBSCRIBE() │
│ + SSUBSCRIBE() / SUNSUBSCRIBE() │
│ + close() / destroy() │
└──────────────┬──────────────────────────┘
│
│ contains
│
┌──────────────────────────┴──────────────────────────┐
│ │
│ RedisClusterSlots │
│ │
│ - slots: Array<Shard>[16384] │
│ - masters: Array<MasterNode> │
│ - replicas: Array<ShardNode> │
│ - nodeByAddress: Map<string, Node> │
│ - pubSubNode?: PubSubNode │
│ - clientSideCache?: PooledClientSideCacheProvider │
│ │
│ Methods: │
│ + connect() │
│ + getClient() │
│ + rediscover() │
│ + getPubSubClient() │
│ + getShardedPubSubClient() │
│ + getRandomNode() │
│ + getSlotRandomNode() │
└───────┬─────────────────┬─────────────────┬─────────┘
│ │ │
┌──────────┘ │ └─────────────┐
│ │ │
│ has many │ optionally has │ has many
▼ ▼ ▼
┌────────────────────────┐ ┌────────────────────────┐ ┌────────────────────────┐
│ Shard │ │ PubSubNode │ │ RedisClient │
│ │ │ │ │ (per node) │
│ - master: MasterNode │ │ - address: string │ │ │
│ - replicas?: Array │ │ - client: RedisClient │ │ Socket, Queue, etc. │
│ - nodesIterator │ │ - connectPromise │ │ │
└──────────┬─────────────┘ └────────────────────────┘ └────────────────────────┘
│
│ contains
│
┌────────────┴────────────┐
│ │
▼ ▼
┌──────────────────┐ ┌──────────────────┐
│ MasterNode │ │ ShardNode │
│ │ │ (replica) │
│ - id: string │ │ │
│ - host: string │ │ - id: string │
│ - port: number │ │ - host: string │
│ - address │ │ - port: number │
│ - readonly: no │ │ - address │
│ - client? │ │ - readonly: yes │
│ - pubSub? │ │ - client? │
│ └─> client │ │ │
│ └─> promise │ │ │
└──────────────────┘ └──────────────────┘


Additional Components:
─────────────────────

┌────────────────────────────────────┐
│ RedisClusterMultiCommand │
│ │
│ Used for MULTI/PIPELINE: │
│ - Batches commands │
│ - Routes to single node │
│ - Returns typed results │
│ │
│ Methods: │
│ + addCommand() │
│ + exec() │
│ + execAsPipeline() │
└────────────────────────────────────┘

┌────────────────────────────────────┐
│ PooledClientSideCacheProvider │
│ (BasicPooledClientSideCache) │
│ │
│ RESP3 Client-Side Caching: │
│ - Shared across all nodes │
│ - Invalidation tracking │
│ - TTL & eviction policies │
│ │
│ Methods: │
│ + get() / set() │
│ + invalidate() │
│ + clear() / enable() / disable() │
└────────────────────────────────────┘


Key Relationships:
─────────────────

1. RedisCluster
└─> RedisClusterSlots (manages topology)
└─> Shard[] (16,384 hash slots)
├─> MasterNode (read/write)
│ └─> RedisClient
│ └─> PubSub RedisClient (sharded pub/sub)
└─> ShardNode[] (replicas, read-only if useReplicas=true)
└─> RedisClient

2. RedisCluster
└─> RedisClusterMultiCommand (for transactions)

3. RedisClusterSlots
└─> PubSubNode (global pub/sub)
└─> RedisClient

4. RedisClusterSlots
└─> PooledClientSideCacheProvider (shared cache, RESP3 only)


Command Flow:
────────────

Single Command:
Client.sendCommand()
→ Cluster._execute()
→ Slots.getClient(key, isReadonly)
→ Calculate slot from key
→ Get Shard for slot
→ Return master or replica client
→ Client.sendCommand()
→ [If MOVED/ASK error]
→ Slots.rediscover()
→ Retry with new node

Transaction (MULTI):
Client.MULTI(routing)
→ RedisClusterMultiCommand
→ Accumulate commands
→ All commands must route to same node
→ client.exec()

Pub/Sub:
Global: Uses single PubSubNode
Sharded: Uses per-master pubSub client based on channel hash


Discovery & Failover:
─────────────────────

1. Initial Connect:
- Try rootNodes in random order
- Execute CLUSTER SLOTS command
- Build slot → shard mapping
- Create client connections

2. Rediscovery (on MOVED error):
- Clear cache
- Re-fetch CLUSTER SLOTS
- Update topology
- Reconnect clients to new nodes

3. Node Address Mapping:
- nodeAddressMap translates cluster IPs
- Useful for NAT/Docker scenarios
Loading
Loading