Learning Ceph - Second Edition电子书

Title Page

Learning Ceph

Second Edition

Copyright

Learning Ceph

Second Edition

Credits

About the Authors

About the Reviewer

www.PacktPub.com

Why subscribe?

Customer Feedback

Preface

What this book covers

What you need for this book

Who this book is for

Conventions

Reader feedback

Customer support

Downloading the example code

Downloading the color images of this book

Errata

Piracy

Questions

Introducing Ceph Storage

The history and evolution of Ceph

Ceph releases

New since the first edition

The future of storage

Ceph as the cloud storage solution

Ceph is software-defined

Ceph is a unified storage solution

The next-generation architecture

RAID: the end of an era

Ceph Block Storage

Ceph compared to other storage solutions

GPFS

iRODS

HDFS

Lustre

Gluster

Ceph

Summary

Ceph Components and Services

Introduction

Core components

Reliable Autonomic Distributed Object Store (RADOS)

MONs

Object Storage Daemons (OSDs)

Ceph manager

RADOS GateWay (RGW)

Admin host

CephFS MetaData server (MDS)

The community

Core services

RADOS Block Device (RBD)

RADOS Gateway (RGW)

CephFS

Librados

Summary

Hardware and Network Selection

Introduction

Hardware selection criteria

Corporate procurement policies

Power requirements-amps, volts, and outlets

Compatibility with management infrastructure

Compatibility with physical infrastructure

Configuring options for one-stop shopping

Memory

RAM capacity and speed

Storage drives

Storage drive capacity

Storage drive form factor

Storage drive durability and speed

Storage drive type

Number of storage drive bays per chassis

Controllers

Storage HBA / controller type

Networking options

Network versus serial versus KVM management

Adapter slots

Processors

CPU socket count

CPU model

Emerging technologies

Summary

Planning Your Deployment

Layout decisions

Convergence: Wisdom or Hype?

Planning Ceph component servers

Rack strategy

Server naming

Architectural decisions

Pool decisions

Replication

Erasure Coding

Placement Group calculations

OSD decisions

Back end: FileStore or BlueStore?

OSD device strategy

Journals

Filesystem

Encryption

Operating system decisions

Kernel and operating system

Ceph packages

Operating system deployment

Time synchronization

Packages

Networking decisions

Summary

Deploying a Virtual Sandbox Cluster

Installing prerequisites for our Sandbox environment

Bootstrapping our Ceph cluster

Deploying our Ceph cluster

Scaling our Ceph cluster

Summary

Operations and Maintenance

Topology

The 40,000 foot view

Drilling down

OSD dump

OSD list

OSD find

CRUSH dump

Pools

Monitors

CephFS

Configuration

Cluster naming and configuration

The Ceph configuration file

Admin sockets

Injection

Configuration management

Scrubs

Logs

MON logs

OSD logs

Debug levels

Common tasks

Installation

Ceph-deploy

Flags

Service management

Systemd: the wave (tsunami?) of the future

Upstart

sysvinit

Component failures

Expansion

Balancing

Upgrades

Working with remote hands

Summary

Monitoring Ceph

Monitoring Ceph clusters

Ceph cluster health

Watching cluster events

Utilizing your cluster

OSD variance and fillage

Cluster status

Cluster authentication

Monitoring Ceph MONs

MON status

MON quorum status

Monitoring Ceph OSDs

OSD tree lookup

OSD statistics

OSD CRUSH map

Monitoring Ceph placement groups

PG states

Monitoring Ceph MDS

Open source dashboards and tools

Kraken

Ceph-dash

Decapod

Rook

Calamari

Ceph-mgr

Prometheus and Grafana

Summary

Ceph Architecture: Under the Hood

Objects

Accessing objects

Placement groups

Setting PGs on pools

PG peering

PG Up and Acting sets

PG states

CRUSH

The CRUSH Hierarchy

CRUSH Lookup

Backfill, Recovery, and Rebalancing

Customizing CRUSH

Ceph pools

Pool operations

Creating and listing pools

Ceph data flow

Erasure coding

Summary

Storage Provisioning with Ceph

Client Services

Ceph Block Device (RADOS Block Device)

Creating and Provisioning RADOS Block Devices

Resizing RADOS Block Devices

RADOS Block Device Snapshots

RADOS Block Device Clones

The Ceph Filesystem (CephFS)

CephFS with Kernel Driver

CephFS with the FUSE Driver

Ceph Object Storage (RADOS Gateway)

Configuration for the RGW Service

Performing S3 Object Operations Using s3cmd

Enabling the Swift API

Performing Object Operations using the Swift API

Summary

Integrating Ceph with OpenStack

Introduction to OpenStack

Nova

Glance

Cinder

Swift

Ganesha / Manila

Horizon

Keystone

The Best Choice for OpenStack storage

Integrating Ceph and OpenStack

Guest Operating System Presentation

Virtual OpenStack Deployment

Summary

Performance and Stability Tuning

Ceph performance overview

Kernel settings

pid_max

kernel.threads-max, vm.max_map_count

XFS filesystem settings

Virtual memory settings

Network settings

Jumbo frames

TCP and network core

iptables and nf_conntrack

Ceph settings

max_open_files

Recovery

OSD and FileStore settings

MON settings

Client settings

Benchmarking

RADOS bench

CBT

FIO

Fill volume, then random 1M writes for 96 hours, no read verification:

Fill volume, then small block writes for 96 hours, no read verification:

Fill volume, then 4k random writes for 96 hours, occasional read verification:

Summary