Главная
Study mode:
on
1
Intro
2
What is this talk about?
3
Terminology
4
Inventory Management
5
Directory Service
6
Automated Provisioning
7
Automated Promotions
8
Central Resource Locking
9
Server's Lifecycle at Facebook
10
Full Disk
11
Bad RAID
12
Old MySQL Version
13
Metadata Database
14
Design Overview
15
Local Triage -MPS' Agent
16
Main Components
17
mps, examples
18
Picking a destination
19
Chosen Algorithm
20
Result: Pretty Graphs
21
Turning Up New Servers
22
Maintenance at Scale
23
The Case of the Helpful Janitor
24
REPLACE ALL THE THINGS!
25
Prepare for takeoff!
26
What happened?
Description:
Explore Facebook's massive MySQL database cluster management in this 50-minute SREcon15 talk. Discover how Facebook automates conventional DBA tasks to operate thousands of servers across multiple data centers. Delve into the design and architecture of their automation systems, including inventory management, automated provisioning, and central resource locking. Learn about server lifecycle management, handling full disks, RAID issues, and MySQL version updates. Examine the metadata database design and local triage processes. Understand how Facebook turns up new servers and manages maintenance at scale. Gain insights from real-world challenges, such as "The Case of the Helpful Janitor" and large-scale replacements. Prepare for an in-depth look at MySQL automation strategies employed by one of the world's largest database clusters.

MySQL Automation at Facebook Scale

USENIX
Add to list