# SOME DESCRIPTIVE TITLE # Copyright (C) YEAR Free Software Foundation, Inc. # This file is distributed under the same license as the PACKAGE package. # FIRST AUTHOR , YEAR. # #, fuzzy msgid "" msgstr "" "Project-Id-Version: PACKAGE VERSION\n" "POT-Creation-Date: 2020-03-02 15:09+0000\n" "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n" "Last-Translator: FULL NAME \n" "Language-Team: LANGUAGE \n" "Language: \n" "MIME-Version: 1.0\n" "Content-Type: text/plain; charset=CHARSET\n" "Content-Transfer-Encoding: 8bit\n" #. type: Title == #: openstack.adoc:2 #, no-wrap msgid "DRBD volumes in Openstack" msgstr "Openstack中的DRBD卷" #. type: Plain text #: openstack.adoc:7 msgid "" "indexterm:[Openstack]indexterm:[Cinder]indexterm:[Nova]In this chapter you " "will learn how to use DRBD in Openstack for persistent, replicated, high-" "performance block storage." msgstr "" "indexterm:[Openstack]indexterm:[Cinder]indexterm:[Nova]在本章中,您将学习如何在Openstack中使用DRBD实现持久的、复制的、高性能的块存储。" #. type: Title === #: openstack.adoc:9 #, no-wrap msgid "Openstack Overview" msgstr "Openstack概述" #. type: Plain text #: openstack.adoc:14 msgid "" "Openstack itself consists of a big range of individual services; the two " "that are mostly concerned with DRBD are Cinder and Nova. *Cinder* is the " "block storage service, while *Nova* is the compute node service that's " "responsible to make the volumes available for the VMs." msgstr "" "Openstack本身包含大量的独立服务;与DRBD相关的两个主要服务是Cinder和Nova。*Cinder* 是块存储服务,*Nova* 是计算节点服务,负责为虚拟机提供卷。" #. type: Plain text #: openstack.adoc:18 msgid "" "DRBD storage volumes can be accessed in two ways: using the iSCSI protocol " "(for maximum compatibility), and using DRBD client functionality (being " "submitted to Openstack). For a discussion of these two modes and their " "differences please see <>." msgstr "" "可以通过两种方式访问DRBD存储卷:使用iSCSI协议(实现最大兼容性)和使用DRBD客户端功能(提交给Openstack)。有关这两种模式及其差异的讨论,请参见<>。" #. type: Title === #: openstack.adoc:20 #, no-wrap msgid "DRBD for Openstack Installation" msgstr "用于Openstack安装的DRBD" #. type: Plain text #: openstack.adoc:24 msgid "" "The _drbdmanage driver_ is upstream in Openstack since the Liberty release. " "It is used (mostly) in the `c-vol` service, so you'll need drbdmanage and " "DRBD 9 installed on the node(s) running that." msgstr "" "自Liberty发布以来,_drbdmanagedriver_ 就已经被合并在Openstack的上游。它主要用于 `c-vol` 服务,因此需要在运行该服务的节点上安装drbdmanage和DRBD" " 9。" #. type: Plain text #: openstack.adoc:27 msgid "" "Depending on the specific Openstack variant being used there are a few " "differences for paths, user names, etc.:" msgstr "根据使用的具体Openstack变量,路径、用户名等有一些不同:" #. type: Block title #: openstack.adoc:28 #, no-wrap msgid "Distribution dependent settings" msgstr "依赖于分发的设置" #. type: Table #: openstack.adoc:37 #, no-wrap msgid "" "what ; rdostack ; devstack\n" "Cinder/DRBD Manage driver file location;\n" "`/usr/lib/python2.6/site-packages/cinder/volume/drivers/drbdmanagedrv.py` ; `/opt/stack/cinder/cinder/volume/drivers/drbdmanagedrv.py`\n" "Cinder configuration file ; `/usr/share/cinder/cinder-dist.conf` ; `/etc/cinder/cinder.conf`\n" "Admin access data, for sourcing into shell ; `/etc/nagios/keystonerc_admin` ; `~stack/devstack/accrc/admin/admin`\n" "User used for running `c-vol` service ; `cinder` ; `stack `\n" msgstr "" "what ; rdostack ; devstack\n" "Cinder/DRBD Manage driver file location;\n" "`/usr/lib/python2.6/site-packages/cinder/volume/drivers/drbdmanagedrv.py` ; `/opt/stack/cinder/cinder/volume/drivers/drbdmanagedrv.py`\n" "Cinder configuration file ; `/usr/share/cinder/cinder-dist.conf` ; `/etc/cinder/cinder.conf`\n" "Admin access data, for sourcing into shell ; `/etc/nagios/keystonerc_admin` ; `~stack/devstack/accrc/admin/admin`\n" "User used for running `c-vol` service ; `cinder` ; `stack `\n" #. type: Plain text #: openstack.adoc:41 msgid "The generalized installations steps are these:" msgstr "一般安装步骤如下:" #. type: Plain text #: openstack.adoc:43 msgid "" "In the `cinder.conf` you'll need something like that; the `volume_driver` " "consists of the class name (last part), and the file path:" msgstr "在 `cinder.conf` 中,您需要类似的东西; `volume_driver` 由类名(最后一部分)和文件路径组成:" #. type: delimited block - #: openstack.adoc:49 #, no-wrap msgid "" "[DEFAULT]\n" "enabled_backends=drbd-1\n" msgstr "" "[DEFAULT]\n" "enabled_backends=drbd-1\n" #. type: delimited block - #: openstack.adoc:54 #, no-wrap msgid "" "[drbd-1]\n" "volume_driver=cinder.volume.drivers.drbdmanagedrv.DrbdManageIscsiDriver\n" "volume_backend_name=DRBD-Managed\n" "drbdmanage_redundancy=1\n" msgstr "" "[drbd-1]\n" "volume_driver=cinder.volume.drivers.drbdmanagedrv.DrbdManageIscsiDriver\n" "volume_backend_name=DRBD-Managed\n" "drbdmanage_redundancy=1\n" #. type: Plain text #: openstack.adoc:59 msgid "" "Please see also <> for choosing between " "iSCSI and DRBD transport modes, and <>." msgstr "" "另请参见<>以选择iSCSI和DRBD传输模式,以及<>。" #. type: Plain text #: openstack.adoc:62 msgid "" "Register the backend: (you might need to fetch the authentication " "environment variables via `source __`)" msgstr "注册后端:(您可能需要通过 `source __` 以获取身份验证环境变量" #. type: delimited block - #: openstack.adoc:68 #, no-wrap msgid "" "# cinder type-create drbd-1\n" "# cinder type-key drbd-1 set volume_backend_name=DRBD-Managed\n" msgstr "" "# cinder type-create drbd-1\n" "# cinder type-key drbd-1 set volume_backend_name=DRBD-Managed\n" #. type: Plain text #: openstack.adoc:72 msgid "" "Allow the user to access the \"__org.drbd.drbdmanaged__\" service on DBus. " "For that you need to extend the file " "`/etc/dbus-1/system.d/org.drbd.drbdmanaged.conf` by an additional stanza " "like this (replace __USER__ by the username as per the table above):" msgstr "" "允许用户访问DBus上的 " "\"__org.drbd.drbdmanaged__\" 服务。为此,您需要将文件 `/etc/dbus-1/system.d/org.drbd.drbdmanaged.conf` 扩展为类似这样的附加节(用上表中的用户名替换用户):" #. type: delimited block - #: openstack.adoc:81 #, no-wrap msgid "" "\n" " \n" " \n" " \n" "\n" msgstr "" "\n" " \n" " \n" " \n" "\n" #. type: Plain text #: openstack.adoc:86 msgid "" "That's it; after a restart of the `c-vol` service you should be able to " "create your DRBD volumes." msgstr "就这样;重启 `c-vol` 服务之后,您应该能够创建您的DRBD卷。" #. type: Title ==== #: openstack.adoc:89 #, no-wrap msgid "Additional Configuration" msgstr "附加配置" #. type: Plain text #: openstack.adoc:93 msgid "" "The `drbdmanage` backend configuration in `cinder.conf` can contain a few " "additional settings that modify the exact behaviour." msgstr "`cinder.conf` 中的 `drbdmanage` 后端配置可以包含一些修改确切行为的附加设置。" #. type: Plain text #: openstack.adoc:99 msgid "" "(((OpenStack,Redundancy))) `drbdmanage_redundancy = 2` e.g. would declare that" " each volume needs to have 2 storage locations, i.e. be replicated once. This" " means that two times the storage will be used, and that the reported free " "space <>." msgstr "" "((OpenStack,Redundancy))`drbdmanage_redundancy=2`例如,将声明每个卷需要有两个存储位置,即复制一次。这意味着将使用两倍的存储空间,并且报告的可用空间看起来是有限的。" #. type: Plain text #: openstack.adoc:104 msgid "" "You can request more than two copies of the data; the limit is given by DRBD" " 9 and the number of storage hosts you have defined." msgstr "您可以请求两个以上的数据副本;限制由DRBD 9和您定义的存储主机数量给出。" #. type: Plain text #: openstack.adoc:109 msgid "" "`drbdmanage_devs_on_controller = True`: By default each volume will get a " "DRBD client mapped on the Cinder controller node; apart from being used for " "iSCSI exports, this might prove helpful for debugging, too." msgstr "" "`drbdmanage_devs_on_controller=True` :默认情况下,每个卷都会在Cinder控制器节点上映射一个DRBD客户机;除了用于iSCSI导出之外,这也可能有助于调试。" #. type: Plain text #: openstack.adoc:113 msgid "" "indexterm:[iSCSI, in OpenStack Cinder]indexterm:[OpenStack, Cinder iSCSI " "transport]In case you need to choose a different iSCSI backend, you can " "provide an additional configuration to set it, like `iscsi_helper=lioadm`." msgstr "" "indexterm:[iSCSI,in OpenStack Cinder]indexterm:[OpenStack,Cinder iSCSI " "transport]如果需要选择不同的iSCSI后端,可以提供其他配置来设置它,如 `iscsi_helper=lioadm`。" #. type: Plain text #: openstack.adoc:117 msgid "" "(((OpenStack,resize policy)))`drbdmanage_resize_policy`, " "(((OpenStack,resource policy)))`drbdmanage_resource_policy`, and " "(((OpenStack,snapshot policy)))`drbdmanage_snapshot_policy` configure" msgstr "" "((OpenStack,resize policy))`drbdmanage_resize_policy`,((OpenStack,resource " "policy))`drbdmanage_resource_policy`,((OpenStack,snapshot " "policy))`drbdmanage_snapshot_policy` 配置" #. type: Plain text #: openstack.adoc:119 msgid "" "the behaviour when resizing volumes, resp. creating snapshots or new " "resources (freshly create or from a snapshot, etc.)" msgstr "重新调整卷大小时的行为。创建快照或新资源(新创建或从快照等)" #. type: Plain text #: openstack.adoc:122 msgid "These are strings that have to be parseable as JSON blobs, for example" msgstr "例如,这些字符串必须可以解析为JSON blob" #. type: Plain text #: openstack.adoc:124 #, no-wrap msgid "\tdrbdmanage_snapshot_policy={'count': '1', 'timeout': '60'}\n" msgstr "\tdrbdmanage_snapshot_policy={'count': '1', 'timeout': '60'}\n" #. type: Plain text #: openstack.adoc:127 msgid "" "See <> for details about the available " "policies and their configuration items." msgstr "有关可用策略及其配置项的详细信息,请参见<>。" #. type: Plain text #: openstack.adoc:132 msgid "" "Please be aware that Python's JSON parser is strict - you'll need to use " "single quotes, for instance, and take other JSON specifications and " "restrictions into account as well!" msgstr "请注意,Python的JSON解析器是严格的 - 例如,您需要使用单引号,并考虑其他JSON规范和限制!" #. type: Plain text #: openstack.adoc:136 msgid "" "In case you want to call different plugins for this purpose, the " "`drbdmanage_resize_plugin`, `drbdmanage_resource_plugin`, and " "`drbdmanage_snapshot_plugin` configuration items exist as well." msgstr "" "如果您想为此目的调用不同的插件,还存在 `drbdmanage_resize_plugin` 、`drbdmanage_resource_plugin` 和 `drbdmanage_snapshot_plugin`配置项。" #. type: Plain text #: openstack.adoc:142 msgid "" "`drbdmanage_net_options` resp. `drbdmanage_resource_options` can be used to " "set DRBD configuration values on each newly created resource. These already " "have sane default values; if you want to override, don't forget to add these" " in again!" msgstr "" "`drbdmanage_net_选项` resp。`drbdmanage_resource_options` 可用于在每个新创建的资源上设置DRBD配置值。这些已经有了正常的默认值;如果要重写,不要忘记再次添加这些值!" #. type: Plain text #: openstack.adoc:145 msgid "Again, these strings get parsed as JSON blobs. The defaults are" msgstr "同样,这些字符串被解析为JSON blob。默认值是" #. type: Plain text #: openstack.adoc:148 #, no-wrap msgid "" " drbdmanage_net_options = {'connect-int': '4', 'allow-two-primaries': 'yes', 'ko-count': '30'}\n" " drbdmanage_resource_options = {'auto-promote-timeout': '300'}\n" msgstr "" " drbdmanage_net_options = {'connect-int': '4', 'allow-two-primaries': 'yes', 'ko-count': '30'}\n" " drbdmanage_resource_options = {'auto-promote-timeout': '300'}\n" #. type: Plain text #: openstack.adoc:155 msgid "" "`drbdmanage_late_local_assign` and `drbdmanage_late_local_assign_exclude` is" " a performance optimization for hyperconverged setups; this needs a bit of " "discussion, so please look at the dedicated chapter <>." msgstr "" "`drbdmanage_late_local_assign`和'drbdmanage_late_local_assign_exclude'是超收敛设置的性能优化;这需要一点讨论,因此请参阅专用章节<>。" #. type: Plain text #: openstack.adoc:158 msgid "" "These configuration settings can be different from one backend to another." msgstr "这些配置设置可以不同于一个后端到另一个后端。" #. type: Title === #: openstack.adoc:162 #, no-wrap msgid "Choosing the Transport Protocol" msgstr "选择传输协议" #. type: Plain text #: openstack.adoc:165 msgid "There are two main ways to run DRBD with Cinder:" msgstr "使用Cinder运行DRBD的方法主要有两种:" #. type: Plain text #: openstack.adoc:167 msgid "accessing storage via <>, and" msgstr "通过<>和" #. type: Plain text #: openstack.adoc:169 msgid "using <> on the wire." msgstr "使用线路上的drbd协议。" #. type: Plain text #: openstack.adoc:172 msgid "" "These are not exclusive; you can define multiple backends, have some of them" " use iSCSI, and others the DRBD protocol." msgstr "这些不是独占的;您可以定义多个后端,其中一些使用iSCSI,另一些使用DRBD协议。" #. type: Title ==== #: openstack.adoc:175 #, no-wrap msgid "iSCSI Transport" msgstr "iSCSI传输" #. type: Plain text #: openstack.adoc:180 msgid "" "The default way to export Cinder volumes is via iSCSI. This brings the " "advantage of maximum compatibility - iSCSI can be used with every " "hypervisor, be it VMWare, Xen, HyperV, or KVM." msgstr "" "导出Cinder卷的默认方法是通过iSCSI。这带来了最大兼容性的优势—无论是VMWare、Xen、HyperV还是KVM,iSCSI都可以用于每个虚拟机监控程序。" #. type: Plain text #: openstack.adoc:184 msgid "" "The drawback is that all data has to be sent to a Cinder node, to be " "processed by an (userspace) iSCSI daemon; that means that the data needs to " "pass the kernel/userspace border, and these transitions will cost some " "performance." msgstr "" "缺点是,所有数据都必须发送到Cinder节点,由(userspace)iSCSI守护进程处理;这意味着数据需要通过内核/userspace边界,这些转换将消耗一些性能。" #. type: Plain text #: openstack.adoc:186 msgid "TODO: performance comparision" msgstr "TODO:性能比较" #. type: Title ==== #: openstack.adoc:189 #, no-wrap msgid "DRBD Transport" msgstr "DRBD传输" #. type: Plain text #: openstack.adoc:196 msgid "" "The alternative is to get the data to the VMs by using DRBD as the transport" " protocol. This means that DRBD 9footnote:[The kernel module and userspace, " "and currently the DRBD Manage daemon too; but please see the note at " "<>.] needs to be installed on the Nova nodes" " too, and so restricts them to Linux with KVM at the moment." msgstr "" "另一种方法是使用DRBD作为传输协议将数据传输到VMs。这意味着DRBD " "9footnote:[内核模块和用户空间,以及当前DRBD管理守护进程;但是请参见<.]中的说明,也需要安装在Nova节点上,因此目前将它们限制为带KVM的Linux。" #. type: Plain text #: openstack.adoc:202 msgid "" "One advantage of that solution is that the storage access requests of the " "VMs can be sent via the DRBD kernel module to the storage nodes, which can " "then directly access the allocated LVs; this means no Kernel/Userspace " "transitions on the data path, and consequently better performance. Combined " "with RDMA capable hardware you should get about the same performance as with" " VMs accessing a FC backend directly." msgstr "" "该解决方案的一个优点是,vm的存储访问请求可以通过DRBD内核模块发送到存储节点,然后存储节点可以直接访问分配的lv;这意味着数据路径上没有内核/用户空间转换,因此性能更好。结合支持RDMA的硬件,您应该获得与VMs直接访问FC后端相同的性能。" #. type: Plain text #: openstack.adoc:206 msgid "" "Another advantage is that you will be implicitly benefitting from the HA " "background of DRBD: using multiple storage nodes, possibly available over " "different network connections, means redundancy and avoiding a single point " "of failure." msgstr "另一个优点是,您将从DRBD的HA背景中隐式获益:使用多个存储节点(可能在不同的网络连接上可用)意味着冗余并避免单点故障。" #. type: Plain text #: openstack.adoc:212 msgid "" "Currently, you'll need to have the hypervisor nodes be part of the DRBD " "Manage cluster." msgstr "目前,您需要让hypervisor节点成为DRBD Manage集群的一部分。" #. type: Plain text #: openstack.adoc:215 msgid "" "When DRBD Manage becomes able to process \"__external nodes__\", the " "requirements on the hypervisor nodes will shrink to DRBD 9 kernel module and" " -userspace only." msgstr "当DRBD Manage能够处理 \"外部节点\" 时,hypervisor节点上的需求将缩小到仅限DRBD 9内核模块和-用户空间。" #. type: Title ==== #: openstack.adoc:219 #, no-wrap msgid "Configuring the Transport Protocol" msgstr "配置传输协议" #. type: Plain text #: openstack.adoc:226 msgid "" "In the storage stanzas in `cinder.conf` you can define the volume driver to " "use; you can use different drivers for different backend configurations, i.e." " you can define a 2-way-redundancy iSCSI backend, a 2-way-redundancy DRBD " "backend, and a 3-way DRBD backend at the same time. Horizonfootnote:[The " "Openstack GUI] should offer these storage backends at volume creation time." msgstr "" "在 `cinder.conf` 中的存储节中,您可以定义要使用的卷驱动程序;您可以为不同的后端配置使用不同的驱动程序,即,您可以同时定义一个双向冗余的iSCSI后端、一个双向冗余的DRBD后端和一个三向DRBD后端。Horizon footnote:[Openstack" " GUI]应该在卷创建时提供这些存储后端。" #. type: Plain text #: openstack.adoc:228 msgid "The available configuration items for the two drivers are" msgstr "两个驱动程序的可用配置项是" #. type: Plain text #: openstack.adoc:230 msgid "for iSCSI:" msgstr "对于iSCSI:" #. type: Plain text #: openstack.adoc:234 #, no-wrap msgid "" "volume_driver=cinder.volume.drivers.drbdmanagedrv.DrbdManageIscsiDriver\n" msgstr "" "volume_driver=cinder.volume.drivers.drbdmanagedrv.DrbdManageIscsiDriver\n" #. type: Plain text #: openstack.adoc:237 msgid "and" msgstr "和" #. type: Plain text #: openstack.adoc:239 msgid "for DRBD:" msgstr "对于DRBD:" #. type: Plain text #: openstack.adoc:243 #, no-wrap msgid "volume_driver=cinder.volume.drivers.drbdmanagedrv.DrbdManageDrbdDriver\n" msgstr "" "volume_driver=cinder.volume.drivers.drbdmanagedrv.DrbdManageDrbdDriver\n" #. type: Plain text #: openstack.adoc:249 msgid "" "The old class name \"DrbdManageDriver\" is being kept for the time because " "of compatibility reasons; it's just an alias to the iSCSI driver." msgstr "由于兼容性原因,旧类名 \"DrbdManageDriver\" 暂时保留;它只是iSCSI驱动程序的别名。" #. type: Title === #: openstack.adoc:252 #, no-wrap msgid "Some further notes" msgstr "一些进一步的说明" #. type: Title ==== #: openstack.adoc:256 #, no-wrap msgid "Free space reporting" msgstr "可用空间报告" #. type: Plain text #: openstack.adoc:260 msgid "" "The free space that the cinder driver reports is fetched from DRBD Manage, " "using the defined <> " "setting." msgstr "" "使用定义的<>设置从DRBD " "Manage获取cinder驱动程序报告的可用空间。" #. type: Plain text #: openstack.adoc:268 msgid "" "This will return the size for the single largest volume that can be created " "with this replication count; so, with 10 storage nodes each having 1TiB free" " space, the value returned for a redundancy count of three will be 1TiB, and" " allocating such a volume will not change the free space value, as there are" " three more nodes with that much free space available. For storage nodes " "with 20GiB, 15GiB, 10GiB, and 5GiB space available, the free space for " "`drbdmanage_redundancy` being 3 will be 10GiB, and 15GiB for 2." msgstr "" "这将返回可使用此复制计数创建的单个最大卷的大小;因此,如果10个存储节点每个都有1TiB可用空间,则为3个冗余计数返回的值将为1TiB,并且分配这样的卷不会更改可用空间值,因为还有三个节点有那么多可用空间。对于具有20GiB、15GiB、10GiB和5GiB可用空间的存储节点, `drbdmanage` 冗余的可用空间为: 3冗余时将为10GiB,而2冗余时将为15GiB。" #. type: Plain text #: openstack.adoc:272 msgid "" "This issue is further muddled by thin LVM pools (one or multiple, depending " "on storage backend in DRBD Manage), and snapshots taken from Cinder volumes." msgstr "瘦LVM池(一个或多个,取决于DRBD Manage中的存储后端)和从Cinder卷中获取的快照进一步混淆了此问题。" #. type: Plain text #: openstack.adoc:274 msgid "" "For further information, please see the Openstack Specs about Thin " "Provisioning" msgstr "有关详细信息,请参阅有关精简配置的Openstack规范" #. type: Plain text #: openstack.adoc:278 msgid "" "there's the https://blueprints.launchpad.net/cinder/+spec/over-subscription-" "in-thin-provisioning[blueprint] and the https://github.com/openstack/cinder-" "specs/blob/master/specs/kilo/over-subscription-in-thin-" "provisioning.rst[text]." msgstr "" "参阅 https://blueprints.launchpad.net/cinder/+spec/over-subscription-" "in-thin-provisioning[blueprint] 和 https://github.com/openstack/cinder-" "specs/blob/master/specs/kilo/over-subscription-in-thin-" "provisioning.rst[text]." #. type: Title ==== #: openstack.adoc:281 #, no-wrap msgid "Hyperconverged Setups" msgstr "超收敛设置" #. type: Plain text #: openstack.adoc:289 msgid "" "The configuration item `drbdmanage_late_local_assign` (available in the DRBD" " Manage Cinder driver from 1.2.0 on, requiring DRBD Manage 0.98.3 or better)" " is a performance optimization for hyperconverged setups. + With that " "feature, the driver tries to get a local copy of the data assigned to the " "hypervisor; that in turn will speed up read IOs, as these won't have to go " "across the network." msgstr "" "配置项 `drbdmanage_late_local_assign`(从1.2.0开始在DRBD Manage Cinder驱动程序中提供,要求DRBD" " Manage " "0.98.3或更高版本)是用于超收敛设置的性能优化。+ 有了这个特性,驱动程序会尝试获取分配给hypervisor的数据的本地副本;这反过来会加快读取IO的速度,因为这些IO不必通过网络。" #. type: Plain text #: openstack.adoc:298 msgid "" "At the time of writing, Nova doesn't pass enough information to Cinder; " "Cinder isn't told which hypervisor will be used. + So the DRBD Manage driver" " assigns all but one copies at `create_volume` time; the last one is done in" " the `attach_volume` step, when the hypervisor is known. If this hypervisor " "is out of space, defined as a storage-less node in DRBD Manage, or otherwise" " not eligible to receive a copy, any other storage node is used instead, and" " the target node will receive a _client_ assignment only." msgstr "" "在编写本文时,Nova没有向Cinder传递足够的信息;Cinder没有被告知将使用哪个hypervisor。+ 因此,DRBD " "Manage驱动程序在 `create_volume` 时间分配除了一个拷贝之外的所有拷贝;最后一个拷贝是在 `attach_volume` 步骤中完成的,这时hypervisor是已知的。如果此虚拟机监控程序空间不足,在DRBD" " Manage中被定义为无存储节点,或者没有资格接收副本,则使用任何其他存储节点,目标节点将仅接收一个客户机分配。" #. type: Plain text #: openstack.adoc:305 msgid "" "Because an image might be copied to the volume before it gets attached to a " "VM, the \"local\" assignment can't simply be done on the first " "accessfootnote:[If it assigned on first access, the image copy node (Glance)" " would receive the copy of the data]. The Cinder driver must be told which " "nodes are not eligible for local copies; this can be done via " "`drbdmanage_late_local_assign_exclude`." msgstr "" "由于映像可能在附加到虚拟机之前复制到卷,因此 \"本地\" 分配不能简单地在第一次访问时完成footprint:[如果在第一次访问时分配,则映像复制节点(Glance)将接收数据的副本]。必须告知Cinder驱动程序哪些节点不符合本地副本的条件;这可以通过 `drbdmanage_late_local_assign_exclude` 完成。" #. type: Plain text #: openstack.adoc:311 msgid "" "For volumes that get cloned from an image stored within Cinder (via a DRBD " "Manage snapshot), the new resource will be empty until the `attach_volume` " "call; at that time the Cinder driver can decide on which nodes the volumes " "will be deployed, and can actually clone the volume on these." msgstr "" "对于从存储在Cinder中的映像(通过DRBD管理快照)中克隆的卷,在调用 `附加卷` 之前,新资源将为空;此时Cinder驱动程序可以决定将部署卷的节点,并可以实际在这些节点上克隆卷。" #. type: Block title #: openstack.adoc:313 #, no-wrap msgid "Free Space Misreported" msgstr "虚报自由空间" #. type: Plain text #: openstack.adoc:319 msgid "" "Late allocation invariably means that the free space numbers are wrong. You " "might prepare 300 VMs, only to find out that you're running out of disk " "space when their volumes are in the middle of synchronizing." msgstr "延迟分配总是意味着可用空间号是错误的。您可能准备了300个虚拟机,但在它们的卷正在同步时发现磁盘空间不足。" #. type: Plain text #: openstack.adoc:322 msgid "" "But that is a common problem with all thin allocation schemes, so we won't " "discuss that in more details here." msgstr "但是这是所有瘦分配方案的一个常见问题,因此我们在这里不再详细讨论这个问题。" #. type: Plain text #: openstack.adoc:326 msgid "To summarize:" msgstr "总结一下:" #. type: Plain text #: openstack.adoc:328 msgid "" "You'll need the DRBD Manage Cinder driver 1.2.0 or later, and DRBD Manage " "0.98.3 or later." msgstr "您将需要DRBD Manage Cinder driver 1.2.0或更高版本,以及DRBD Manage 0.98.3或更高版本。" #. type: Plain text #: openstack.adoc:330 msgid "" "The <> must be used; iSCSI won't " "offer any locality benefits." msgstr "必须使用<>;iSCSI不会提供任何本地好处。" #. type: Plain text #: openstack.adoc:332 msgid "" "The <> must be set " "to at least two copies." msgstr "" "必须将<>设置为至少两个副本。" #. type: Plain text #: openstack.adoc:334 msgid "" "To generally enable this feature, set `drbdmanage_late_local_assign` to " "`True`." msgstr "要一般启用此功能,请将 `drbdmanage_late_local_assign` 设置为 `True` 。" #. type: Plain text #: openstack.adoc:336 msgid "" "To specify which hosts should *not* get a local copy, set " "`drbdmanage_late_local_assign_exclude` to a comma-separated list of " "hostnames; this should typically include Glance and the Cinder-controller " "nodes (but not the Cinder-storage nodes!)." msgstr "" "若要指定哪些主机 *不应* 获取本地副本,请将 `drbdmanage_late_local_assign_exclude` 设置为以逗号分隔的主机名列表;这通常应包括Glance和Cinder控制器节点(但不包括Cinder存储节点!)。" #. type: Plain text #: openstack.adoc:338 msgid "Take care to not run out of disk space." msgstr "注意不要用完磁盘空间。" #. There is a performance comparison in preparion, showing iSCSI/local/DRBD #. hyperconverged #. type: Plain text #: openstack.adoc:345 msgid "Here are a few links that show you collected performance data." msgstr "这里有几个链接,显示您收集的性能数据。" #. type: Plain text #: openstack.adoc:349 msgid "" "https://www.3ware.co.jp[Thirdware Inc.] did a Ceph vs. DRBD9 comparison, " "too; the japanese original can be found in their " "https://www.3ware.co.jp/download/technical-docs[technical documentation] " "area." msgstr "" "https://www.3ware.co.jp[Thirdware " "Inc.]也对Ceph和DRBD9进行了比较;日语原版可以在其 https://www.3ware.co.jp/download/technical " "docs[technical documentation]区域找到。" #. type: Plain text #: openstack.adoc:350 msgid "" "A translated (English) version is available on request at sales@linbit.com." msgstr "如有要求,可通过sales@linbit.com获得翻译(英文)版本。" #. type: Plain text #: openstack.adoc:352 msgid "http://links.linbit.com/Ceph-DRBD9[\"__Ceph vs. DRBD9 Performance" msgstr "http://links.linbit.com/Ceph-DRBD9[\"__Ceph vs. DRBD9性能" #. type: Plain text #: openstack.adoc:354 msgid "" "Comparison__\"] discusses IOPs, bandwidth, and IO latency; this one needs a " "free registration on the LINBIT site." msgstr "\"]讨论了IOPs、带宽和IO延迟;这需要在LINBIT站点上进行免费注册。"