[openstack swift] Deployment Guide部署指南

================

Deployment Guide部署指南

================

 

-----------------------

Hardware Considerations硬件因素

-----------------------

 

Swift is designed to run on commodity hardware. At Rackspace, our storage

servers are currently running fairly generic 4U servers with 24 2T SATA

drives and 8 cores of processing power. RAID on the storage drives is not

required and not recommended. Swift's disk usage pattern is the worst

case possible for RAID, and performance degrades very quickly using RAID 5

or 6.

Swift被设计运行于商用硬件之上。在Rackspace, 我们的存储服务器目前都是有24个2T SATA硬盘和

8核处理器的4U服务器。不需要也不推荐使用RAID硬盘。Swift的硬盘使用方式对RAID盘来说是最不适合的，

而且对RAID 6和6,会出现性能的急剧下降。

------------------

Deployment Options部署选项

------------------

 

The swift services run completely autonomously, which provides for a lot of

flexibility when architecting the hardware deployment for swift. The 4 main

services are:

Swift的服务都是自动运行的，它们在部署swift时提供极佳的灵活性。四项服务是：

 

#. Proxy Services

#. Object Services

#. Container Services

#. Account Services

 

The Proxy Services are more CPU and network I/O intensive. If you are using

10g networking to the proxy, or are terminating SSL traffic at the proxy,

greater CPU power will be required.

 

Proxy服务是CPU和网络I/O密集型。如果你给Proxy提供10网络，或者打算结束SSL阻塞的问题，你需要更

强悍的CPU。

 

The Object, Container, and Account Services (Storage Services) are more disk

and network I/O intensive.

 

Object，Container和Account服务则需要更多的存储空间，同时也是网络I/O密集型。

 

The easiest deployment is to install all services on each server. There is

nothing wrong with doing this, as it scales each service out horizontally.

 

最容易的部署方案是将以上所有的服务都安装到每台服务器。这样做没有什么不对的，因为这样可以水平地扩展

每个服务。

 

At Rackspace, we put the Proxy Services on their own servers and all of the

Storage Services on the same server. This allows us to send 10g networking to

the proxy and 1g to the storage servers, and keep load balancing to the

proxies more manageable.  Storage Services scale out horizontally as storage

servers are added, and we can scale overall API throughput by adding more

Proxies.

 

在Rackspace,我们将Proxy Services以及另外3个存储服务都部署在每台服务器上。这样可以使我们分配

10G的带宽给Proxy,分配1G给存储服务器，同时在所有代理服务间保持更可控的负载均衡。当当存储服务器增

加时，存储服务就会水平扩展，并且我们可以通过增加Proxies来拓展接口吞吐。

 

If you need more throughput to either Account or Container Services, they may

each be deployed to their own servers. For example you might use faster (but

more expensive) SAS or even SSD drives to get faster disk I/O to the databases.

 

如果你需要使Account和Container服务有更大的吞吐量，你需要将它们独立部署到专门服务器上。

比如你需要使用更快速同时也更昂贵的SAS或者甚至SSD盘来提高提高数据库的I/O性能。

 

Load balancing and network design is left as an excercise to the reader,

but this is a very important part of the cluster, so time should be spent

designing the network for a Swift cluster.

 

负载均衡和网络设计就作为给读者留下的练习吧，但是它们也是非常重要的部分，因此你极有必要花费一些时间

在此问题上。

 

.. _ring-preparing:

 

------------------

Preparing the Ring准备环

------------------

 

The first step is to determine the number of partitions that will be in the

ring. We recommend that there be a minimum of 100 partitions per drive to

insure even distribution accross the drives. A good starting point might be

to figure out the maximum number of drives the cluster will contain, and then

multiply by 100, and then round up to the nearest power of two.

 

第一步是确定环中分区的数量。我们建议每个磁盘有至少100个分区以确保对磁盘分布的均匀。

最好先弄清楚集群将会包含的磁盘的总数，然后乘以100，然后取最接近此值的2的幂。

 

For example, imagine we are building a cluster that will have no more than

5,000 drives. That would mean that we would have a total number of 500,000

partitions, which is pretty close to 2^19, rounded up.

 

比如，想象我们正在构建一个将有多于5000个磁盘的集群。这将意味着我们应有500，000个分区，那最接近

的2的幂就是2的19次方。

 

It is also a good idea to keep the number of partitions small (realatively).

The more partitions there are, the more work that has to be done by the

replicators and other backend jobs and the more memory the rings consume in

process. The goal is to find a good balance between small rings and maximum

cluster size.

 

将分区数量保持在一个小一点的值也是个很好的主意(比较而言)。分区数量越多，在复制和其它一些后台操作

时就需要做更多的工作，环也将消耗更多的内存。

 

The next step is to determine the number of replicas to store of the data.

Currently it is recommended to use 3 (as this is the only value that has

been tested). The higher the number, the more storage that is used but the

less likely you are to lose data.

 

第二个步骤是确定存储数据副本的数量。目前建议存3份(这也是目前唯一经过测试的值)。当副本数量越多时，

就需要越多的存粗空间，越少时，数据丢失的可能就更大。

 

It is also important to determine how many zones the cluster should have. It is

recommended to start with a minimum of 5 zones. You can start with fewer, but

our testing has shown that having at least five zones is optimal when failures

occur. We also recommend trying to configure the zones as high a level as

possible to create as much isolation as possible. Some example things to take

into consideration can include physical location, power availability, and

network connectivity. For example, in a small cluster you might decide to

split the zones up by cabinet, with each cabinet having its own power and

network connectivity. The zone concept is very abstract, so feel free to use

it in whatever way best isolates your data from failure. Zones are referenced

by number, beginning with 1.

 

确定一个集群有几个zone也是非常重要的。建议一开始至少分5个区域。当然你可以一开始设置更少，但是我们

的测试显示在错误发生时，最少5个区域的方案是最优的。我们也建议常识设置更多的区域并且将它们保持尽可能

高的独立性。比如物理位置，能源状况和网络状况这些因素可以在分区时加入参考。比如在一个小集群中你可以

以机柜为单位来分区，每一个机柜享有独立的供电和独立的网络连接。区域的概念是很抽象的，所以只要能使你

的每份数据拥有更高的隔离性，那么请不要有任何顾虑地使用它。区域使用数字来引用，以1开始。

 

You can now start building the ring with::

你现在可以通过一下命令来构建环

 

    swift-ring_builder <builder_file> create <part_power> <replicas> <min_part_hours>

 

This will start the ring build process creating the <builder_file> with 

2^<part_power> partitions. <min_part_hours> is the time in hours before a

specific partition can be moved in succession (24 is a good value for this).

 

这将启动环构造进程来生成包含2^<part_power>个分区的builder_file。<min_part_hours>是指定分区

可以被成功移动的间隔时间（最好是24小时）。

 

Devices can be added to the ring with::

存储设备可以通过一下命令来添加到环中

 

    swift-ring-builder <builder_file> add z<zone>-<ip>:<port>/<device_name>_<meta> <weight>

 

This will add a device to the ring where <builder_file> is the name of the

builder file that was created previously, <zone> is the number of the zone

this device is in, <ip> is the ip address of the server the device is in,

<port> is the port number that the server is running on, <device_name> is

the name of the device on the server (for example: sdb1), <meta> is a string

of metadata for the device (optional), and <weight> is a float weight that

determines how many partitions are put on the device relative to the rest of

the devices in the cluster (a good starting point is 100.0 x TB on the drive).

Add each device that will be initially in the cluster.

 

这条命令会将存储设备添加到与前面创建的同名的那个builder_file的环文件中。<zone>是此设备所在的

区域号码，<ip>是设备所在服务器的ip,<port>是对应的端口，<device_name>是设备在服务器中的设备

名(比如sdb1),<meta>是可选的设备元数据字符串，<weight>是存储设备的浮点类型的权重，此权重与该

设备比集群中其它设备得到的分区数量成正比(好的起点是设备都是100T的倍数）,以上将添加到集群中的设

备初始化。

 

Once all of the devices are added to the ring, run::

一旦所有设备都添加到环上了，运行：

    swift-ring-builder <builder_file> rebalance

 

This will distribute the partitions across the drives in the ring. It is

important whenever making changes to the ring to make all the changes

required before running rebalance. This will ensure that the ring stays as

balanced as possible, and as few partitions are moved as possible.

 

这会使设备对应的分区均匀地分布在环上。在运行均衡命令之前就应该修改完毕所有需要的地方。这将确保环

尽可能地均衡，和更少的分区会被移动。

 

The above process should be done to make a ring for each storage serivce

(Account, Container and Object). The builder files will be needed in future

changes to the ring, so it is very important that these be kept and backed up.

The resulting .tar.gz ring file should be pushed to all of the servers in the

cluster. For more information about building rings, running

swift-ring-builder with no options will display help text with available

commands and options. More information on how the ring works internally

can be found in the :doc:`Ring Overview <overview_ring>`.

 

以上操作需要对每个存储服务都执行一遍以生成对应的环。环的构建文件会在未来对环的修改时用到，因此

保存好且备份此文件至关重要。.tar.gz结尾的环文件应该被推送到集群中的所有服务器上。欲得到更多关于

构建环的信息，不带参数地运行swift-ring会得到帮助文件。关于环内部是如何工作的更多信息，请参阅

Ring Overview <overview_ring>文档。

 

---------------------------

Object Server Configuration对象服务器配置

---------------------------

 

An Example Object Server configuration can be found at 

etc/object-server.conf-sample in the source code repository.

 

etc/object-server.conf-sample是对象服务器的范本。

 

The following configuration options are available:

以下配置选项是有效的：

 

[object-server]

object服务器

 

==================  ==========  =============================================

Option              Default     Description

------------------  ----------  ---------------------------------------------

swift_dir           /etc/swift  Swift configuration directory

Swift配置文件目录

devices             /srv/node   Parent directory of where devices are mounted

devices挂载点的父目录

mount_check         true        Weather or not check if the devices are

                                mounted to prevent accidently writing

                                to the root device

是否检查devices被以防止偶然写入root设备的方式挂载

bind_ip             0.0.0.0     IP Address for server to bind to

object server的IP

bind_port           6000        Port for server to bind to

object server的端口

workers             1           Number of workers to fork

启动几个工作实例

log_facility        LOG_LOCAL0  Syslog log facility

日志记录设备

log_level           INFO        Logging level

日志记录级别

log_requests        True        Weather or not to log each request

是否对每个请求记录日志

user                swift       User to run as

以何用户运行

node_timeout        3           Request timeout to external services

外部服务的请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务的连接超时时间

network_chunk_size  65536       Size of chunks to read/write over the

                                network

网络读写块的大小

disk_chunk_size     65536       Size of chunks to read/write to disk

磁盘读写块的大小

max_upload_time     86400       Maximum time allowed to upload an object

允许上传object的最大时间。

slow                0           If > 0, Minimum time in seconds for a PUT

                                or DELETE request to complete

如果大于0，PUT或DELETE请求完成的最小时间（单位：秒）

==================  ==========  =============================================

 

[object-replicator]

object复制器

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

daemonize           yes         Weather or not to run replication as a

                                daemon

是否以守护进程的形式运行复制进程

run_pause           30          Time in seconds to wait between replication

                                passes

复制进程过时的时间（单位：秒）

concurrency         1           Number of replication workers to spawn

复制操作实例并发数量

timeout             5           Timeout value sent to rsync --timeout and

                                --contimeout options

发送给rsync超时时间

stats_interval      3600        Interval in seconds between logging

                                replication statistics

记录复制操作统计日志的间隔时间

reclaim_age         604800      Time elapsed in seconds before an object

                                can be reclaimed

object可被回收的时间

==================  ==========  ===========================================

 

[object-updater]

objecy更新器

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

interval            300         Minimum time for a pass to take

使上一次审计得以完成的最小时间间隔

concurrency         1           Number of updater workers to spawn

更新器实例并发数量

node_timeout        10          Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

slowdown            0.01        Time in seconds to wait between objects

objects间等待时间

==================  ==========  ===========================================

 

[object-auditor]

object审计师

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

interval            1800        Minimum time for a pass to take

使上一次审计得以完成的最小时间间隔

node_timeout        10          Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

==================  ==========  ===========================================

 

------------------------------

Container Server Configuration容器服务器配置

------------------------------

 

An example Container Server configuration can be found at 

etc/container-server.conf-sample in the source code repository.

 

etc/container-server.conf-sample是容器服务器配置的范本。

 

The following configuration options are available:

 

[container-server]

容器服务器

==================  ==========  ============================================

Option              Default     Description

------------------  ----------  --------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

swift_dir           /etc/swift  Swift configuration directory

Swift配置文件目录

devices             /srv/node   Parent directory of where devices are mounted

devices挂载点的父目录

mount_check         true        Weather or not check if the devices are

                                mounted to prevent accidently writing

                                to the root device

是否检查devices是否以防止偶然写入root设备的方式挂载

bind_ip             0.0.0.0     IP Address for server to bind to

container server的IP

bind_port           6001        Port for server to bind to

container server的端口

workers             1           Number of workers to fork

contaienr server的工作实例

user                swift       User to run as

以何账户运行

node_timeout        3           Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

==================  ==========  ============================================

 

[container-replicator]

container复制器

 

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

per_diff            1000

concurrency         8           Number of replication workers to spawn

生成几个复制器工作实例

run_pause           30          Time in seconds to wait between replication

                                passes

复制进程过时的时间（单位：秒）

node_timeout        10          Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

reclaim_age         604800      Time elapsed in seconds before a container

                                can be reclaimed

container可被回收的时间

==================  ==========  ===========================================

 

[container-updater]

container更新器

 

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

interval            300         Minimum time for a pass to take

使上一次审计得以完成的最小时间间隔

concurrency         4           Number of updater workers to spawn

产生更新器工作实例的数量

node_timeout        3           Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

slowdown            0.01        Time in seconds to wait between containers

containers间等待时间

==================  ==========  ===========================================

 

[container-auditor]

container审计师

 

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

interval            1800        Minimum time for a pass to take

使上一次审计得以完成的最小时间间隔

node_timeout        10          Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

==================  ==========  ===========================================

 

----------------------------

Account Server Configuration帐号服务器配置

----------------------------

 

An example Account Server configuration can be found at 

etc/account-server.conf-sample in the source code repository.

 

etc/account-server.conf-sample是帐号服务器范本

 

The following configuration options are available:

以下配置选项是有效的

 

[account-server]

account服务器

==================  ==========  =============================================

Option              Default     Description

------------------  ----------  ---------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

swift_dir           /etc/swift  Swift configuration directory

Swift配置文件目录

devices             /srv/node   Parent directory or where devices are mounted

devices挂载点的父目录

mount_check         true        Weather or not check if the devices are

                                mounted to prevent accidently writing

                                to the root device

是否检查devices是否以防止偶然写入root设备的方式挂载

bind_ip             0.0.0.0     IP Address for server to bind to

帐号服务器的IP

bind_port           6002        Port for server to bind to

帐号服务器的端口

workers             1           Number of workers to fork

帐号服务器工作实例的个数

user                swift       User to run as

以何帐号运行

==================  ==========  =============================================

 

[account-replicator]

account复制器

 

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

per_diff            1000

concurrency         8           Number of replication workers to spawn

复制器工作实例并发个数

run_pause           30          Time in seconds to wait between replication

                                passes

复制进程过时时间

node_timeout        10          Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

reclaim_age         604800      Time elapsed in seconds before a account

                                can be reclaimed

account可被回收的时间

==================  ==========  ===========================================

 

[account-auditor]

account审计师

 

====================  ==========  ===========================================

Option                Default     Description

--------------------  ----------  -------------------------------------------

log_facility          LOG_LOCAL0  Syslog log facility

   系统日志记录设备

log_level             INFO        Logging level

   日志记录级别

interval              1800        Minimum time for a pass to take

   使上一次审计得以完成的最小时间间隔

max_container_count   100         Maximum containers randomly picked for

                                  a given account audit

 随机抽查每个account的最多containers个数

node_timeout          10          Request timeout to external services

   外部服务请求超时时间

conn_timeout          0.5         Connection timeout to external services

   外部服务连接超时时间

====================  ==========  ===========================================

 

[account-reaper]

account回收器

==================  ==========  ===========================================

Option              Default     Description

------------------  ----------  -------------------------------------------

log_facility        LOG_LOCAL0  Syslog log facility

系统日志记录设备

log_level           INFO        Logging level

日志记录级别

concurrency         25          Number of replication workers to spawn

回收期工作实例并发数量

interval            3600        Minimum time for a pass to take

使上一次回收得以完成的最小时间间隔

node_timeout        10          Request timeout to external services

外部服务请求超时时间

conn_timeout        0.5         Connection timeout to external services

外部服务连接超时时间

==================  ==========  ===========================================

 

--------------------------

Proxy Server Configuration代理服务器配置

--------------------------

 

[proxy-server]

代理服务器

 

============================  ===============  =============================

Option                        Default          Description

----------------------------  ---------------  -----------------------------

log_facility                  LOG_LOCAL0       Syslog log facility

    系统日志记录设备

log_level                     INFO             Log level

    日志记录级别

bind_ip                       0.0.0.0          IP Address for server to

                                               bind to

  服务器IP

bind_port                     80               Port for server to bind to

    服务器端口

cert_file                                      Path to the ssl .crt

  到ssl .crt的路径 

key_file                                       Path to the ssl .key

  到ssl .key的路径

swift_dir                     /etc/swift       Swift configuration directory

    Swift配置文件目录

log_headers                   True             If True, log headers in each

                                               request

  如果为真，则记录每一个请求的头信息

workers                       1                Number of workers to fork

    服务器工作实例的个数

user                          swift            User to run as

    以何用户运行

recheck_account_existence     60               Cache timeout in seconds to

                                               send memcached for account

                                               existance

  重新为已存账户发送内存缓存的缓存超时时间

recheck_container_existence   60               Cache timeout in seconds to

                                               send memcached for container

                                               existance

  重新为已存容器发送内存缓存的缓存超时时间

object_chunk_size             65536            Chunk size to read from

                                               object servers

  从对象服务器读取数据的块大小

client_chunk_size             65536            Chunk size to read from

                                               clients

  从clients端读取数据的块大小

memcache_servers              127.0.0.1:11211  Comma separated list of

                                               memcached servers ip:port

  memcache_servers的ip和port

node_timeout                  10               Request timeout to external

                                               services

  外部服务请求超时时间

client_timeout                60               Timeout to read one chunk

                                               from a client

  从一个client读取一个数据块的超时时间

conn_timeout                  0.5              Connection timeout to

                                               external services

  外部服务连接超时时间

error_suppression_interval    60               Time in seconds that must

                                               elapse since the last error

                                               for a node to be considered

                                               no longer error limited

  从最后一次报出error起，间隔多久会被认为已经正常

error_suppression_limit       10               Error count to consider a

                                               node error limited

  建一个node判定为出错的的error个数

rate_limit                    20000.0          Max container level ops per

                                               second

  每秒最大级别的容器操作

account_rate_limit            200.0            Max account level ops per

                                               second

  每秒最大级别的账户操作

rate_limit_account_whitelist                   Comma separated list of 

                                               account name hashes to not

                                               rate limit

  

rate_limit_account_blacklist                   Comma separated list of

                                               account name hashes to block

                                               completly

============================  ===============  =============================

 

[auth-server]

 

============  ===================================  ========================

Option        Default                              Description

------------  -----------------------------------  ------------------------

class         swift.common.auth.DevAuthMiddleware  Auth wsgi middleware

                                                   to use

  使用的Auth wsgi中间件

ip            127.0.0.1                            IP address of auth

                                                   server

  认证服务器的IP

port          11000                                Port of auth server

    认证服务器的端口

node_timeout  10                                   Request timeout

    请求超时时间

============  ===================================  ========================

 

------------------------

Memcached Considerations Memcached因素

------------------------

 

Several of the Services rely on Memcached for caching certain types of

lookups, such as auth tokens, and container/account existance.  Swift does

not do any caching of actual object data.  Memcached should be able to run

on any servers that have available RAM and CPU.  At Rackspace, we run 

Memcached on the proxy servers.  The `memcache_servers` config option

in the `proxy-server.conf` should contain all memcached servers.

 

好几个服务都依赖Memcached来缓存几种查找项，比如校验口令，以及container/account是否存在。

Swift不对对象数据进行任何缓存。Memcached应该可以运行在任何有满足需求的RAM和CPU的服务器上。

在Rackspace,我们将Memcached存储在Proxy Servers上。

'proxy-server.conf'的'memcache_servers'配置选项需要包含所有的缓存服务器。

 

-----------

System Time系统时间

-----------

 

Time may be relative but it is relatively important for Swift!  Swift uses

timestamps to determine which is the most recent version of an object.

It is very important for the system time on each server in the cluster to

by synced as closely as possible (more so for the proxy server, but in general

it is a good idea for all the servers).  At Rackspace, we use NTP with a local

NTP server to ensure that the system times are as close as possible.  This

should also be monitored to ensure that the times do not vary too much.

 

时间或许是相对而言的，但是它却是Swift中非常重要的一个因素！Swift使用时间戳来确定哪个才是对象的

最新版本。将集群中的所有服务器的系统时间同步到最同步的状态非常重要(尤其是Proxy,但是最好将所有服

务器的时间都同步)。在Rackspace,我们使用本地的一个NTP服务器来确保系统的时间尽可能地接近。这同样

也应该被严密监视以确保系统时间相差不会太大。

 

----------------------

General Service Tuning一般服务调节

----------------------

 

Most services support either a worker or concurrency value in the settings.

This allows the services to make effective use of the cores available. A good

starting point to set the concurrency level for the proxy and storage services

to 2 times the number of cores available. If more than one service is

sharing a server, then some experimentaiton may be needed to find the best

balance.

 

大多数的服务都支持一个或者多个并发的处理器。这使服务可以更大化地利用多核的性能。一个好的起点是将

Proxy和存储服务器的并发水平设置为内核数量的两倍。如果有１个以上的服务在共享一个服务器，那需要做

些具体的实验以找出最平衡的方式。

 

At Rackspace, our Proxy servers have dual quad core processors, giving us 8

cores. Our testing has shown 16 workers to be a pretty good balance when

saturating a 10g network and gives good CPU utilization.

 

在Rackspace,我们的Proxy服务器都有两个４核的处理器，这为我们提供了８个核。

我们的测试显示１６个实例会最大化地利用10g网络带宽和CPU能力。

 

Our Storage servers all run together on the same servers. These servers have

dual quad core processors, for 8 cores total. We run the Account, Container,

and Object servers with 8 workers each. Most of the background jobs are run

at a concurrency of 1, with the exception of the replicators which are run at

a concurrency of 2.

 

我们的存储服务都运行在同样的服务器上：２＊４核，公８核。对Account,Container和对象服务我们均

运行８个实例。多数的后台处理都运行一个并发的实例上，而复制器则运行２个并发的实例。

 

The above configuration setting should be taken as suggestions and testing

of configuration settings should be done to ensure best utilization of CPU,

network connectivity, and disk I/O.

 

以上的配置都应该依照建议进行，同时需要测试以寻找最优方案来实现CPU,网络连接，磁盘I/O的最有效利用。

-------------------------

Filesystem Considerations文件系统因素

-------------------------

 

Swift is designed to be mostly filesystem agnostic--the only requirement

beeing that the filesystem supports extended attributes (xattrs). After

thorough testing with our use cases and hardware configurations, XFS was

the best all-around choice. If you decide to use a filesystem other than

XFS, we highly recommend thorough testing.

 

Swift被设计可适用大多数的文件系统－－－唯一的要求是该文件系统必须支持扩展属性(xattrs)。我们的

使用案例和硬件配置的测试显示，XFS 是综合性能最优的。如果你决定使用其它的文件系统，我们强烈建议

进行测试。

 

If you are using XFS, some settings that can dramatically impact

performance. We recommend the following when creating the partition::

 

如果你在使用XFS,一些配置会显著地影响性能。我们建议在创建分区时执行一下命令：

 

    mkfs.xfs -i size=1024 -f /dev/sda1

 

Setting the inode size is important, as XFS stores xattr data in the inode.

If the metadata is too large to fit in the inode, a new extent is created,

which can cause quite a performance problem. Upping the inode size to 1024

bytes provides enough room to write the default metadata, plus a little

headroom. We do not recommend running Swift on RAID, but if you are using

RAID it is also important to make sure that the proper sunit and swidth

settings get set so that XFS can make most efficient use of the RAID array.

 

设置inode的size是非常重要的，因为XFS存储xattr数据在inode上。如果元数据太大而inode不足以容纳，

那么一个扩展就会被创建，这会导致性能问题。将inode大小提升值1024bytes就可以提供足够空间以写入默认

元数据，并增加一点headroom.我们不建议在RAID上运行Swift,如果你在使用RAID，那么确保合适的sunit

和swidth配置也非常重要，这样XFS可以更有效使用RAID array。

 

We also recommend the following example mount options when using XFS::

我们建议在使用XFS时以下列方式来mount:

 

    mount -t xfs -o noatime,nodiratime,nobarrier,logbufs=8 /dev/sda1 /srv/node/sda

 

For a standard swift install, all data drives are mounted directly under

/srv/node (as can be seen in the above example of mounting /def/sda1 as

/srv/node/sda). If you choose to mount the drives in another directory,

be sure to set the `devices` config option in all of the server configs to

point to the correct directory.  

 

对一个标准的Swift安装，所有的数据磁盘都被直接挂载在/srv/node下(如上会将/dev/sda1挂载在/srv/node/sda)。

如果你选择将磁盘挂载在别的目录，请确保将设备所有的设备配置都指向正确的目录。

---------------------

General System Tuning一般系统调节

---------------------

 

Rackspace currently runs Swift on Ubuntu Server 10.04, and the following

changes have been found to be useful for our use cases.

 

Rackspace目前运行Swift在Ubuntu Server 10.04上，以下的一些修改已经被发现对我们的用例来说是

非常有用的：

 

The following settings should be in `/etc/sysctl.conf`::

 

    # disable TIME_WAIT.. wait..

    net.ipv4.tcp_tw_recycle=1

    net.ipv4.tcp_tw_reuse=1

 

    # disable syn cookies

    net.ipv4.tcp_syncookies = 0

 

    # double amount of allowed conntrack

    net.ipv4.netfilter.ip_conntrack_max = 262144

 

To load the updated sysctl settings, run ``sudo sysctl -p``

 

要重新载入更新的sysctl设置，请运行"sudo sysctl -p"

 

A note about changing the TIME_WAIT values.  By default the OS will hold

a port open for 60 seconds to ensure that any remaining packets can be

received.  During high usage, and with the number of connections that are

created, it is easy to run out of ports.  We can change this since we are

in control of the network.  If you are not in control of the network, or

do not expect high loads, then you may not want to adjust those values.

 

修改TIME_WAIT时需要注意。操作系统默认情况下会保持一个６０秒的端口以确保任何包都可以被接收到。

在高度使用中，并且在创建那么多数量的连接时，端口会很容易就被耗尽。我们可以改变这个因为我们控制

了网络。如果你未能控制网络，或者不期望高的负载，你可能不想调节这些值。

 

----------------------

Logging Considerations日志因素

----------------------

 

Swift is set up to log directly to syslog. Every service can be configured

with the `log_facility` option to set the syslog log facility destination.

It is recommended to use syslog-ng to route the logs to specific log

files locally on the server and also to remote log collecting servers.

 

Swift会直接将日志记录在syslog中。每个服务都可以通过配置`log_facility`来设置日志记录的目的地。

建议使用syslog-ng来将日志封发到指定的本地日志文件中，同时也封发到远程的日志收集服务器。