Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Scaling Your Data Safely with Riak
Search
Alex Moore
January 10, 2014
Programming
3
140
Scaling Your Data Safely with Riak
CodeMash 2014 Presentation.
Alex Moore
January 10, 2014
Tweet
Share
Other Decks in Programming
See All in Programming
『リコリス・リコイル』に学ぶ!! 〜キャリア戦略における計画的偶発性理論と変わる勇気の重要性〜
wanko_it
1
610
自作OSでDOOMを動かしてみた
zakki0925224
1
1.4k
エンジニアのための”最低限いい感じ”デザイン入門
shunshobon
0
130
20250808_AIAgent勉強会_ClaudeCodeデータ分析の実運用〜競馬を題材に回収率100%の先を目指すメソッドとは〜
kkakeru
0
210
サイトを作ったらNFCタグキーホルダーを爆速で作れ!
yuukis
0
650
Introduction to Git & GitHub
latte72
0
120
GUI操作LLMの最新動向: UI-TARSと関連論文紹介
kfujikawa
0
1k
Microsoft Orleans, Daprのアクターモデルを使い効率的に開発、デプロイを行うためのSekibanの試行錯誤 / Sekiban: Exploring Efficient Development and Deployment with Microsoft Orleans and Dapr Actor Models
tomohisa
0
220
SOCI Index Manifest v2が出たので調べてみた / Introduction to SOCI Index Manifest v2
tkikuc
1
110
画像コンペでのベースラインモデルの育て方
tattaka
3
1.9k
TDD 実践ミニトーク
contour_gara
1
200
「リーダーは意思決定する人」って本当?~ 学びを現場で活かす、リーダー4ヶ月目の試行錯誤 ~
marina1017
0
250
Featured
See All Featured
The Straight Up "How To Draw Better" Workshop
denniskardys
236
140k
XXLCSS - How to scale CSS and keep your sanity
sugarenia
248
1.3M
Producing Creativity
orderedlist
PRO
347
40k
Put a Button on it: Removing Barriers to Going Fast.
kastner
60
4k
Faster Mobile Websites
deanohume
309
31k
A better future with KSS
kneath
239
17k
Visualizing Your Data: Incorporating Mongo into Loggly Infrastructure
mongodb
48
9.7k
Keith and Marios Guide to Fast Websites
keithpitt
411
22k
Improving Core Web Vitals using Speculation Rules API
sergeychernyshev
18
1.1k
The MySQL Ecosystem @ GitHub 2015
samlambert
251
13k
Designing for Performance
lara
610
69k
4 Signs Your Business is Dying
shpigford
184
22k
Transcript
with Riak @alexmoore Scaling Data Safely
Internet 1 (2) (3)
Internet 503 Timeout 1 (2) (3)
Internet 1 (2) 3
Internet 1 (2) 3
Internet (2) 3 2 3 1
Internet 4 2 3 (2) 1 3 3
Internet CDN 2 6 7 2 4 3 1 3
3
Internet CDN 5 6 2 4 3 2 1 3
3 7 3
Internet CDN 6 7 2 3 2 1 3 3
Internet CDN P S S 3 6 7 2 5
3 2 1 4 4 3 3 7
Internet CDN 6 8 2 3 2 1 3 3
Internet CDN 6 8 2 3 2 1 3 3
3? 4? q User Session Widgets Lolcats Stuff 8? magic
Internet CDN 6 8 2 3 2 1 3 3
3? 4? q User Session Widgets Lolcats Stuff 8? magic
Internet CDN 6 8 2 3 2 1 3 3
3? 4? q User Session Widgets Lolcats Stuff 8? magic
Internet CDN 6 8 2 3 2 1 3 3
3? 4? q User Session Widgets Lolcats Stuff 8? magic Who’s on first?
Internet CDN 6 8 2 3 2 1 3 3
3? 4? q User Session Widgets Lolcats Stuff 8? magic What the hell have you built? Who’s on first?
Enter Riak
Riak is a distributed, fault-tolerant, highly available database
None
Distributed
Distributed a cluster of nodes each node has an identical
role
Fault- tolerant
Fault- Tolerant symmetrical no Single Point of Failure
Highly Available
Highly Available replicas & redundancy
Scalable
Scalable add or remove nodes linear scalability
Scalable add or remove nodes linear scalability
Multiple Platforms
Multiple Platforms
Multiple Platforms
Predictable Performance
None
Internet CDN 5 6 2 3 2 1 3 3
Internet CDN 5 6 2 3 2 1 3 3
3 3 4 4
1 2 3 4 5
6 7 8 9 10 11 12 Building A Cluster # All Nodes $ riak start ! # Nodes 2..n.each $ riak-‐admin join
[email protected]
! # Any Node $ riak-‐admin cluster plan $ riak-‐admin cluster commit !
None
None
Architecture
Riak is the ops-friendly database
Riak is distributed
Consistent Hashing hash(“bucket/key”)
hash ring
tokenize it
node 0 node 1 node 2
node 0 node 1 node 2 hash(key)
node 0 node 1 node 2 Replicas are stored to
the N - 1 contiguous partitions
node 0 node 1 node 2 hash(“companies/GE”) Replicas are stored
to the N - 1 contiguous partitions
node 0 node 1 node 2 hash(“companies/GE”) Replicas are stored
to the N - 1 contiguous partitions
node 0 node 1 node 2
node 0 node 1 node 2 node 3 + Scaling
out
Quorum requests N R W PR/PW DW
get(“users/clay-davis”)
get(“users/clay-davis”) client Riak
get(“users/clay-davis”) Get Handler (FSM) client Riak
get(“users/clay-davis”) Get Handler (FSM) client Riak hash(“users/clay-davis”) == 10, 11,
12
get(“users/clay-davis”) Get Handler (FSM) client Riak hash(“users/clay-davis”) == 10, 11,
12 Coordinating node Cluster 6 7 8 9 10 11 12 13 14 15 16 The Ring
get(“users/clay-davis”) Get Handler (FSM) client Riak get(“users/clay-davis”) Coordinating node Cluster
6 7 8 9 10 11 12 13 14 15 16 The Ring
get(“users/clay-davis”) Get Handler (FSM) client Riak Coordinating node Cluster 6
7 8 9 10 11 12 13 14 15 16 The Ring R=2
get(“users/clay-davis”) Get Handler (FSM) client Riak Coordinating node Cluster 6
7 8 9 10 11 12 13 14 15 16 The Ring R=2 obj
get(“users/clay-davis”) Get Handler (FSM) client Riak R=2 obj obj
get(“users/clay-davis”) Get Handler (FSM) client Riak R=2 obj obj
get(“users/clay-davis”) obj
Read Repair (Anti-Entropy)
replica replica replica
replica replica replica X
replica replica replica replica replica replica
Active Anti-Entropy (self healing clusters)
real-time updates persistent non-blocking disk-based
merkle tree to track changes coordinated at the vnode level
runs as a background process exchange with neighbor vnodes for inconsistencies resolution semantics: trigger read-repair
= hashes marked “dirty”
None
None
None
None
= keys to read-repair
Using Riak
Riak is the ops-friendly database
Riak is friendly to Developers, too!
{“key”: “value”}
key value key value key value key value key value
key value key value key value key value Keys are namespaced into Buckets
Buckets aren’t “real”
✘ ✘ ✘ Buckets aren’t “real”
Metadata Bucket Properties
Aitch Tee Tee Pee GET /buckets/bucket/keys/key PUT /buckets/bucket/keys/key DEL /buckets/bucket/keys/key
1 2 3 4 5
6 7 8 9 10 11 12 Working With Ruby - Starting $ gem install riak-‐client ! book = { isbn: '1111979723', title: 'Moby Dick', author: 'Herman Melville', body: 'Call me Ishmael. Some years ago...', copies_owned: 3 }
1 2 3 4 5
6 7 8 9 10 11 12 Working With Ruby - Create # Starting Client client = Riak::Client.new(protocol: 'pbc', pb_port: 8087, host: '10.0.0.1') ! books_bucket = client.bucket('books') ! robj_moby = books_bucket.new(book[:isbn]) robj_moby.data = book robj_moby.store ! !
1 2 3 4 5
6 7 8 9 10 11 12 Working With Ruby - Read fetched_book = books_bucket.get('1111979723') ! puts fetched_book.raw_data ! => {"isbn":"1111979723","title":"Moby Dick”, "author":"Herman Melville", "body":"Call me Ishmael. Some years ago...","copies_owned":3} ! fetched_book.data['isbn'] ! => "1111979723"
1 2 3 4 5
6 7 8 9 10 11 12 Working With Ruby - Delete fetched_book.delete ! # or ! books_bucket.delete('1111979723')
1 2 3 4 5
6 7 8 9 10 11 12 Working With Ruby - 2I ! fetched_book.indexes['author_bin'] = [] fetched_book.indexes['author_bin'] << fetched_book.data['author'] ! fetched_book.store ! books_bucket.get_index('author_bin','Herman Melville') ! => ["1111979723"]
etc...
JUST ONE MORE THING
Internet CDN 5 6 2 3 2 1 3 3
3 3 4 4
Internet CDN 5 6 2 3 2 1 3 3
3 4
Internet CDN 5 6 2 3 2 1 3 3
3 4
in riak.conf:
search = off in riak.conf:
search = off search = on in riak.conf:
1 2 3 4 5
6 7 8 9 10 11 12 client.create_search_index(“tweets_idx”) ! bucket = Riak::Bucket.new(client, "tweets") bucket.props = {"search_index" => "tweets_idx"} ! # load some tweets into riak… ! results = client.search("tweets", "text:who") ! puts results[‘docs'].first
1 2 3 4 5
6 7 8 9 10 11 12 pp results[‘docs'].first ! {"score"=>"2.50804900000000019489e+00", "_yz_rb"=>"tweets", "_yz_rk"=>"421672897250226176", ...}
More resources: docs.basho.com “Taste of Riak”
Multi-Datacenter Replication (E.G. - Nuke proofing)
Full Sync
Real-Time
None
put(“users/jimmy”)
put(“users/jimmy”)
put(“users/jimmy”)
put(“users/jimmy”)
put(“users/jimmy”)
Internet CDN 5 6 2 3 2 1 3 3
3 4
Internet CDN 5 6 2 3 2 1 3 3
3 4 ?
Internet CDN 5 6 2 3 2 1 3 3
3 4 ?
Q & A Alex Moore - Client Services Engineer, Basho
@alexmoore
[email protected]
http://alexmoore.io