本文详细介绍了FastDFS的组成、上传机制、使用场景及架构，并提供了一步一步的部署教程，包括安装依赖、配置Tracker和Storage服务器、Nginx集成以及模拟文件迁移。FastDFS是一款轻量级分布式文件系统，适用于互联网应用存储中小文件。文章还涵盖了FastDFS的常用命令和故障排查。
FastDFS 介绍及安装部署
https://github.com/happyfish100
FastDFS 是一款开源的轻量级分布式文件系统，
- 纯C实现，支持Linux、FreeBSD等类UNIX系统；
- 类Google FS，不是通用的文件系统，只能通过专有的API访问，目前提供了 C、Java和PHP API；
- 为互联网应用量身定做，追求高性能和高扩展型；
- FastDFS可以看做是基于文件系统的key/value pair存储系统，称作分布式文件存储服务更为合适；
- 更倾向于存储中小型文件（4KB~500MB）；
FastDFS 组成
Tracker Server
跟踪服务器，主要做调度工作，在访问时起负载均衡的作用。在内存中记录 集群中group和storage服务器的状态信息，是连接客户端和Storage服务端的枢纽。因为相关信息全部在内存中，Tracker服务器的性能非常高，一个较大的集群（比如上百个group）中有3台就足够了。
Storage Server
存储服务器，文件和文件属性（meta data）都保存到存储服务器上。
FastDFS上传机制
title FastDFS上传机制
participant Client
participant Tracker_Server
participant Storage_Server
Storage_Server -> Tracker_Server: 1. 定时向Tracker上传状态信息
Client -> Tracker_Server: 2. 上传连接请求
Tracker_Server -> Tracker_Server: 3. 查询可用storage
Client <- Tracker_Server: 4. 返回信息(storage的IP和端口)
Client -> Storage_Server: 5. 上传文件(file content和meta data)
Storage_Server -> Storage_Server: 6. 生成file_id
Storage_Server -> Storage_Server: 7. 将上传内容写入硬盘
Client <- Storage_Server: 8. 返回file_id(路径信息和文件名)
Client -> Client: 9. 存储文件信息
FastDFS使用场景
FastDFS架构
FastDFS 架构图如下：
实验环境
| 主机名 | IP地址 | 说明 |
|---|---|---|
| tracker01.lavenliu.com | 192.168.20.160 | 追踪服务器01 |
| tracker02.lavenliu.com | 192.168.20.161 | 追踪服务器02 |
| storage01.lavenliu.com | 192.168.20.162 | 存储服务器01 |
| storage02.lavenliu.com | 192.168.20.163 | 存储服务器02 |
| client01.lavenliu.com | 192.168.20.164 | 客户端01 |
上述表作废，以下面的为准，所有服务安装在一台机器上。
| 主机名 | IP地址 | 说明 |
|---|---|---|
| tracker.lavenliu.com | 192.168.16.137 | 追踪服务器 |
| storage.lavenliu.com | 192.168.16.137 | 存储服务器 |
| client.lavenliu.com | 192.168.16.137 | 客户端 |
| xxx | 10.20.20.35 | 迁移用服务器 |
实验环境中用到的版本信息：
- Nginx 1.21.1
- FastDFS 6.0.7
- FastDFS libfastcommon 1.0.48
- Nginx FastDFS Module 1.22
部署 FastDFS
安装依赖
官方文档如下：
# step 1. download libfastcommon source codes and install it,
# github address: https://github.com/happyfish100/libfastcommon.git
# gitee address: https://gitee.com/fastdfs100/libfastcommon.git
# command lines as:
git clone https://github.com/happyfish100/libfastcommon.git
cd libfastcommon; git checkout V1.0.47
./make.sh clean && ./make.sh && ./make.sh install
# step 2. download fastdfs source codes and install it,
# github address: https://github.com/happyfish100/fastdfs.git
# gitee address: https://gitee.com/fastdfs100/fastdfs.git
# command lines as:
git clone https://github.com/happyfish100/fastdfs.git
cd fastdfs; git checkout V6.07
./make.sh clean && ./make.sh && ./make.sh install
# step 3. setup the config files
# the setup script does NOT overwrite existing config files,
# please feel free to execute this script (take easy :)
./setup.sh /etc/fdfs
# step 4. edit or modify the config files of tracker, storage and client
such as:
vi /etc/fdfs/tracker.conf
vi /etc/fdfs/storage.conf
vi /etc/fdfs/client.conf
and so on ...
# step 5. run the server programs
# start the tracker server:
/usr/bin/fdfs_trackerd /etc/fdfs/tracker.conf restart
# start the storage server:
/usr/bin/fdfs_storaged /etc/fdfs/storage.conf restart
# (optional) in Linux, you can start fdfs_trackerd and fdfs_storaged as a service:
/sbin/service fdfs_trackerd restart
/sbin/service fdfs_storaged restart
# step 6. (optional) run monitor program
# such as:
/usr/bin/fdfs_monitor /etc/fdfs/client.conf
# step 7. (optional) run the test program
# such as:
/usr/bin/fdfs_test <client_conf_filename> <operation>
/usr/bin/fdfs_test1 <client_conf_filename> <operation>
# for example, upload a file for test:
/usr/bin/fdfs_test /etc/fdfs/client.conf upload /usr/include/stdlib.h
tracker server config file sample please see conf/tracker.conf
storage server config file sample please see conf/storage.conf
client config file sample please see conf/client.conf
Item detail
1. server common items
---------------------------------------------------
| item name | type | default | Must |
---------------------------------------------------
| base_path | string | | Y |
---------------------------------------------------
| disabled | boolean| false | N |
---------------------------------------------------
| bind_addr | string | | N |
---------------------------------------------------
| network_timeout | int | 30(s) | N |
---------------------------------------------------
| max_connections | int | 256 | N |
---------------------------------------------------
| log_level | string | info | N |
---------------------------------------------------
| run_by_group | string | | N |
---------------------------------------------------
| run_by_user | string | | N |
---------------------------------------------------
| allow_hosts | string | * | N |
---------------------------------------------------
| sync_log_buff_interval| int | 10(s) | N |
---------------------------------------------------
| thread_stack_size | string | 1M | N |
---------------------------------------------------
memo:
* base_path is the base path of sub dirs:
data and logs. base_path must exist and it's sub dirs will
be automatically created if not exist.
$base_path/data: store data files
$base_path/logs: store log files
* log_level is the standard log level as syslog, case insensitive
# emerg: for emergency
# alert
# crit: for critical
# error
# warn: for warning
# notice
# info
# debug
* allow_hosts can ocur more than once, host can be hostname or ip address,
"*" means match all ip addresses, can use range like this: 10.0.1.[1-15,20]
or host[01-08,20-25].domain.com, for example:
allow_hosts=10.0.1.[1-15,20]
allow_hosts=host[01-08,20-25].domain.com
2. tracker server items
---------------------------------------------------
| item name | type | default | Must |
---------------------------------------------------
| port | int | 22000 | N |
---------------------------------------------------
| store_lookup | int | 0 | N |
---------------------------------------------------
| store_group | string | | N |
---------------------------------------------------
| store_server | int | 0 | N |
---------------------------------------------------
| store_path | int | 0 | N |
---------------------------------------------------
| download_server | int | 0 | N |
---------------------------------------------------
| reserved_storage_space| string | 1GB | N |
---------------------------------------------------
memo:
* the value of store_lookup is:
0: round robin (default)
1: specify group
2: load balance (supported since V1.1)
* store_group is the name of group to store files.
when store_lookup set to 1(specify group),
store_group must be set to a specified group name.
* reserved_storage_space is the reserved storage space for system
or other applications. if the free(available) space of any stoarge
server in a group <= reserved_storage_space, no file can be uploaded
to this group (since V1.1)
bytes unit can be one of follows:
# G or g for gigabyte(GB)
# M or m for megabyte(MB)
# K or k for kilobyte(KB)
# no unit for byte(B)
3. storage server items
-------------------------------------------------
| item name | type | default | Must |
-------------------------------------------------
| group_name | string | | Y |
-------------------------------------------------
| tracker_server | string | | Y |
-------------------------------------------------
| port | int | 23000 | N |
-------------------------------------------------
| heart_beat_interval | int | 30(s) | N |
-------------------------------------------------
| stat_report_interval| int | 300(s) | N |
-------------------------------------------------
| sync_wait_msec | int | 100(ms) | N |
-------------------------------------------------
| sync_interval | int | 0(ms) | N |
-------------------------------------------------
| sync_start_time | string | 00:00 | N |
-------------------------------------------------
| sync_end_time | string | 23:59 | N |
-------------------------------------------------
| store_path_count | int | 1 | N |
-------------------------------------------------
| store_path0 | string |base_path| N |
-------------------------------------------------
| store_path# | string | | N |
-------------------------------------------------
|subdir_count_per_path| int | 256 | N |
-------------------------------------------------
|check_file_duplicate | boolean| 0 | N |
-------------------------------------------------
| key_namespace | string | | N |
-------------------------------------------------
| keep_alive | boolean| 0 | N |
-------------------------------------------------
| sync_binlog_buff_interval| int | 60s | N |
-------------------------------------------------
memo:
* tracker_server can ocur more than once, and tracker_server format is
"host:port", host can be hostname or ip address.
* store_path#, # for digital, based 0
* check_file_duplicate: when set to true, must work with FastDHT server,
more detail please see INSTALL of FastDHT. FastDHT download page:
http://code.google.com/p/fastdht/downloads/list
* key_namespace: FastDHT key namespace, can't be empty when
check_file_duplicate is true. the key namespace should short as possible
简化的步骤：
git clone https://gitee.com/fastdfs100/libfastcommon.git
cd libfastcommon
git checkout V1.0.48
./make.sh clean && ./make.sh && ./make.sh install
安装服务端
git clone https://gitee.com/fastdfs100/fastdfs.git
cd fastdfs
git checkout V6.07
./make.sh clean
./make.sh
[root@node01 fastdfs]# ./make.sh install
mkdir -p /usr/bin
mkdir -p /etc/fdfs
cp -f fdfs_trackerd /usr/bin
if [ ! -f /etc/fdfs/tracker.conf ]; then cp -f ../conf/tracker.conf /etc/fdfs/tracker.conf; fi
if [ ! -f /etc/fdfs/storage_ids.conf ]; then cp -f ../conf/storage_ids.conf /etc/fdfs/storage_ids.conf; fi
mkdir -p /usr/bin
mkdir -p /etc/fdfs
cp -f fdfs_storaged /usr/bin
if [ ! -f /etc/fdfs/storage.conf ]; then cp -f ../conf/storage.conf /etc/fdfs/storage.conf; fi
mkdir -p /usr/bin
mkdir -p /etc/fdfs
mkdir -p /usr/lib64
mkdir -p /usr/lib
cp -f fdfs_monitor fdfs_test fdfs_test1 fdfs_crc32 fdfs_upload_file fdfs_download_file fdfs_delete_file fdfs_file_info fdfs_appender_test fdfs_appender_test1 fdfs_append_file fdfs_upload_appender fdfs_regenerate_filename /usr/bin
if
扫一扫
7956
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
