StarlingX系统具备升级的能力,这个特性叫作“patching”,提供从2个版本之间升级的能力,主要用户bug修复、安全补丁和特性加强等等。node
Patching支持两种补丁, In-Service补丁和Reboot-required补丁。In-Service补丁不须要主机节点重启,只须要服务进程重启便可。Reboot-required补丁须要重启主机以实现补丁生效。在升级Reboot-required补丁时,须要先对主机进行lock操做,等待补丁applied,再unlock使其生效。api
这篇介绍文档,主要面向开发人员使用补丁功能,而不是产品用户指南。它更着重介绍补丁修复功能,而不是包含补丁的各个方面。安全
简要的说,补丁修复包含2个阶段,建立补丁和应用补丁。下面进行详细介绍这两部分。app
一个StarlingX补丁包括一个或多个系统升级所须要的rpm包。在开始建立补丁以前须要验证rpm包已经在已部署的StarlingX上安装了。如下步骤能够帮助咱们确认。async
肯定已部署系统的软件版本,有两种方式
• 在horizon界面上 Admin -> Platform
-> System Configuration -> Systems
• 使用命令行 system show
controller-0:~$ . /etc/platform/openrc
[sysadmin@controller-0 ~(keystone_admin)]$ system show
+----------------------+--------------------------------------+
| Property | Value |
+----------------------+--------------------------------------+
| contact | None |
| created_at | 2019-10-14T03:10:50.862114+00:00 |
| description | None |
| https_enabled | False |
| location | None |
| name | 608dfe48-9a05-4b21-afc1-ea122574caa7 |
| region_name | RegionOne |
| sdn_enabled | False |
| security_feature | spectre_meltdown_v1 |
| service_project_name | services |
| software_version | 19.09 |
| system_mode | duplex |
| system_type | All-in-one |
| timezone | UTC |
| updated_at | 2019-10-14T03:12:41.983029+00:00 |
| uuid | 2639ad15-08a7-4f1b-a372-f927a5e4ab31 |
| vswitch_type | none |
+----------------------+--------------------------------------+ide
如今假设StarlingX的源码已经下载好了,须要升级安装的rpm包也准备好了,如今咱们开始构造补丁构建环境。再次声明,这个教程主要针对开发人员,而不是产品。工具
安装2个cgcs-patch依赖包,crypto和pycrypto
sudo pip install crypto pycryptoui
在这个脚本中,它从release-info.inc 文件中获取PLATFORM_RELEASE参数,并把PYTHONPATH指向repo中的cgcs-patch包,避免了安装cgcs-patch和手动指定PLATFORM_RELEASE参数。可使用下面命令查看构建脚本的使用说明。this
$ $MY_REPO/stx/stx-update/cgcs-patch/bin/patch_build --help Usage: patch_build [ <args> ] ... <rpm list> Options: --id <id> Patch ID --release <version> Platform release version --status <status> Patch Status Code (ie. O, R, V) --unremovable Marks patch as unremovable --reboot-required <Y|N> Marks patch as reboot-required (default=Y) --summary <summary> Patch Summary --desc <description> Patch Description --warn <warnings> Patch Warnings --inst <instructions> Patch Install Instructions --req <patch_id> Required Patch --controller <rpm> New package for controller --worker <rpm> New package for worker node --worker-lowlatency <rpm> New package for worker-lowlatency node --storage <rpm> New package for storage node --controller-worker <rpm> New package for combined node --controller-worker-lowlatency <rpm> New package for lowlatency combined node --all-nodes <rpm> New package for all node types
使用这个脚本能够指定patch id、reboot required、depended patches、rpm list等等,若是系统上没有的,须要新安装的包须要指定节点,好比 --controller 指定是在控制节点上新装包。脚本执行完后,能够获得名字为“<patch-id>.patch”的文件。spa
下面深刻研究下这个补丁文件。
首先,这个补丁文件是个gzip压缩包。咱们能够经过file命令去检查下。
$ file 001.patch
001.patch: gzip compressed data, was "001.patch", last modified:
Fri Aug 16 05:56:59 2019, max compression
解压出来后,能够看到如下文件
$ tar -xf 001.patch
$ tree
├── 001.patch
├── metadata.tar
├── signature
├── signature.v2
└── software.tar
$MY_REPO/build-tools/signing/ima_signing_key.priv
在metadata.tar中只有一个文件metadata.xml,包含补丁构建的全部信息。StarlingX集群系统会读取这个文件信息。
signature文件包含software.tar和metadata.tar的MD5的组合。
补丁生成后,能够手动安装补丁到指定的StarlingX系统,同时支持界面和命令行安装操做。补丁的生命周期包括四个状态: Available,Partial-Apply, Applied 和 Partial-Remove.
• Available:表示补丁已经上传到补丁存储库里了,可是尚未到软件升级的仓库,一样也没有在任何主机上安装。
• Partial-Apply:表示补丁升级程序已经被触发(sw-patch apply),已经在部分主机上安装,可是尚未在全部须要安装的主机上安装。
• Applied:表示已经在全部须要安装的主机上安装完成。
• Partial-Remove:表示补丁正在被移除,经过命令触发(sw-patch remove),正在移除,可是还没彻底移除。
若是须要用命令行安装补丁,须要把补丁拷贝到active的控制节点上。StarlingX集群提供客户端命令sw-patch。补丁操做都是经过这个命令完成,这个命令提供了不少功能,包括upload, apply, query,host-install, delete, remove等等。
controller-0:~$ sw-patch --help usage: sw-patch [--debug] <subcommand> ... Subcomands: upload: Upload one or more patches to the patching system. upload-dir: Upload patches from one or more directories to the patching system. apply: Apply one or more patches. This adds the specified patches to the repository, making the update(s) available to the hosts in the system. Use --all to apply all available patches. Patches are specified as a space-separated list of patch IDs. remove: Remove one or more patches. This removes the specified patches from the repository. Patches are specified as a space-separated list of patch IDs. delete: Delete one or more patches from the patching system. Patches are specified as a space-separated list of patch IDs. query: Query system patches. Optionally, specify 'query applied' to query only those patches that are applied, or 'query available' to query those that are not. show: Show details for specified patches. what-requires: List patches that require the specified patches. query-hosts: Query patch states for hosts in the system. host-install: Trigger patch install/remove on specified host. To force install on unlocked node, use the --force option. host-install-async: Trigger patch install/remove on specified host. To force install on unlocked node, use the --force option. Note: This command returns immediately upon dispatching installation request. install-local: Trigger patch install/remove on the local host. This command can only be used for patch installation prior to initial configuration. drop-host: Drop specified host from table. query-dependencies: List dependencies for specified patch. Use --recursive for recursive query. is-applied: Query Applied state for list of patches. Returns True if all are Applied, False otherwise. report-app-dependencies: Report application patch dependencies, specifying application name with --app option, plus a list of patches. Reported dependencies can be dropped by specifying app with no patch list. query-app-dependencies: Display set of reported application patch dependencies. commit: Commit patches to free disk space. WARNING: This action is irreversible! --os-region-name: Send the request to a specified region
下面演示如何使用这个命令去安装补丁。演示用的补丁是须要安装在全部主机上的In-Service补丁,须要升级的StarlingX环境是 2+2+2的标准环境。
Patch Current 表示当前主机是否有补丁安装,Yes表示没有安装补丁,No表示至少有一个补丁在安装
当补丁状态available后,能够触发补丁安装
controller-0:/$ sudo sw-patch apply 001
001 is now in the repo
检查补丁状态
controller-0:~$ sudo sw-patch query
Patch ID RR Release Patch State
======== == ======= =============
001 N 19.09 Partial-Apply
检查节点状态
controller-0:~$ sudo sw-patch query-hosts
Hostname IP Address Patch Current Reboot Required Release State
============ ============== ============= =============== ====== =====
compute-0 192.178.204.7 No No 19.09 idle
compute-1 192.178.204.9 No No 19.09 idle
controller-0 192.178.204.3 No No 19.09 idle
controller-1 192.178.204.4 No No 19.09 idle
storage-0 192.178.204.12 No No 19.09 idle
storage-1 192.178.204.11 No No 19.09 idle
在每一个节点上安装补丁,因为是in-service 补丁,因此不须要执行lock操做。
controller-0:~$ sudo sw-patch host-install controller-0
...
Installation was successful.
检查主机升级状态
controller-0:~$ sudo sw-patch query-hosts
Hostname IP Address Patch Current Reboot Required Release State
============ ============== ============= =============== ====== =====
compute-0 192.178.204.7 No No 19.09 idle
compute-1 192.178.204.9 No No 19.09 idle
controller-0 192.178.204.3 Yes No 19.09 idle
controller-1 192.178.204.4 No No 19.09 idle
storage-0 192.178.204.12 No No 19.09 idle
storage-1 192.178.204.11 No No 19.09 idle
在全部节点上安装补丁,须要为每一个节点执行命令
controller-0:~$ sudo sw-patch host-install controller-1
....
Installation was successful.
controller-0:~$ sudo sw-patch host-install compute-0
....
Installation was successful.
controller-0:~$ sudo sw-patch host-install compute-1
....
Installation was successful.
controller-0:~$ sudo sw-patch host-install storage-0
...
Installation was successful.
controller-0:~$ sudo sw-patch host-install storage-1
...
Installation was successful.
除了补丁升级,StarlingX还支持补丁回退和删除,经过下面两个命令实现sw-patch remove和sw-patch host-install,和补丁安装有点相似。
在上面的例子中,演示了在集群中补丁升级的功能。可是在大规模集群中,整个升级过程会持续很长的时间。特别是reboot-required补丁,这个方案会很糟糕,效率很低并且给管理员带来不少工做。所以StarlingX提供了另外一个高级特性“补丁编排”。它支持集群经过一些简单的操做达到升级的目的,极大减小管理员的工做负担和较少出错。这个功能有三种方式使用,客户端CLI、界面Horizon和VIM Restful API。
客户端CLI。StarlingX提供客户端工具sw-manager, 能够用于补丁编排。以下所示,能够经过建立和应用补丁策略来升级整个集群
controller-0:~$ sw-manager patch-strategy -h
usage: sw-manager patch-strategy [-h] ...
optional arguments:
-h, --help show this help message and exit
Software Patch Commands:
create Create a strategy
delete Delete a strategy
apply Apply a strategy
abort Abort a strategy
show Show a strategy
controller-0:~$ sw-manager patch-strategy create -h
usage: sw-manager patch-strategy create [-h]
[--controller-apply-type {serial,ignore}]
[--storage-apply-type {serial,parallel,ignore}]
[--worker-apply-type {serial,parallel,ignore}]
[--max-parallel-worker-hosts {2,3,4,5,6,7,8,9,10,
11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,
28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,
45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,
62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,
79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,
96,97,98,99,100}]
[--instance-action {migrate,stop-start}]
[--alarm-restrictions {strict,relaxed}]
optional arguments:
-h, --help show this help message and exit
--controller-apply-type {serial,ignore}
defaults to serial
--storage-apply-type {serial,parallel,ignore}
defaults to serial
--worker-apply-type {serial,parallel,ignore}
defaults to serial
--max-parallel-worker-hosts {2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,
17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,
37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,
57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,
77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,
97,98,99,100}
maximum worker hosts to patch in parallel
--instance-action {migrate,stop-start}
defaults to stop-start
--alarm-restrictions {strict,relaxed}
defaults to strict
界面Horizon。打开Admin -> Platform
-> Software Management -> Patch Orchestration 标签
在补丁安装时,补丁编排要求集群处于一个良好的状态。
• 全部主机必须处于unlocked-enabled-available状态
• 系统没有告警
• 足够的空间用于VM迁移
• 全部的源码都在StarlingX仓库里开源,包括“update”和“nfv”• in-service补丁和reboot-required补丁的生成和安装已经通过验证• 补丁编排还没通过验证