用velero和nutanix对象备份karbon kubernetes群集

2年前
2019年10月20日
0答复
2447意见

UserLevel 4

乔森纳克斯
Nutanix员工
78个答复

免责声明：该帖子旨在用于演示目的，不得将其视为准备就绪。Nutanix Karbon不包括Velero，因此Nutanix的支持将无法处理与Velero有关的任何情况。

概述

集装箱应用程序的主要原理之一是无状态。原因不是使此应用程序可移植和不依赖任何数据。通过这种方式，您可以在任何平台上重新使用相同的容器映像，并具有相同的结果。

由于容器开始流行，因为可移植性，可伸缩性等，社区找到了一种使用本地存储或共享量的诸如数据库（数据库）（例如数据库）的方法。

恢复无状态应用程序是一个非常简单的过程，您只需要将清单文件重新应用于另一个集群即可。您只需要确保您的清单文件是最新的，并且在群集中运行的最新状态即可。请记住，您不应该直接在群集中进行更改，而应使用新更改更新清单文件，并将其应用于群集。

那州申请呢？好吧，由于使用容器，数据与应用程序解耦，因此备份过程可能有些挑战。有本机命令将您的持续卷的文件系统复制到另一个位置，或使用基础存储的本机功能将数据复制到DR站点。但是，当您想在DR站点中恢复数据时，此过程需要其他任务，并且还取决于您使用的Kubernetes存储插件，您可能还需要导出秘密键才能将存储存储安装在DR中地点。

为了简化无状态和状态应用程序的备份和还原过程，Heptio开发了ARK，现在称为Velero。在此博客中，我们将使用Velero在带有Kubernetes的Nutanix Karbon群集中备份Zookeeper应用程序。备份将被推到Nutanix对象桶中。

htDRtfSVrVv1TTUgqQE2C7XJ92qtGByhxW17Tgoq_J1jDmkI9s5__cFcwO8E_MSzza8LuU1BUv3GyfOqdb_FnBletrVGgDYGI2sAVutjVu0_cwb2iCdqRaJYneJ5SvpqBkb4RPQg

什么是Velero？

[资源：velero github repo这是给予的

Velero（以前是Heptio ARK）为您提供了备份和恢复Kubernetes群集资源和持续数量的工具。您可以使用云提供商或本地运行Velero。Velero让您：

在损失的情况下备份群集并恢复。
将集群资源迁移到其他群集。
将您的生产集群复制到开发和测试簇。

Velero由：

在群集上运行的服务器
当地运行的命令行客户端

Velero支持使用称为Restic的免费开源备份工具来备份和恢复Kubernetes量。这种支持被认为是β质量。请参阅清单限制要了解当前是否适合您的用例。

什么是restic？

[资源：Restic GitHub仓库这是给予的

Restic是一个快速，高效且安全的备份程序。它支持三个主要操作系统（Linux，MacOS，Windows）和一些较小的操作系统（FreeBSD，OpenBSD）。

什么是nutanix对象？

nutanix对象是一种软件定义的对象存储解决方案，可在降低整体成本的同时不间断地扩展。它采用与S3兼容的REST API接口设计，可将Terabytes处理到非结构化数据的pet，全部来自一个单个名称空间。对象专为备份，长期保留/档案和跨区域DevOps团队而设计。它是Nutanix Enterprise Cloud平台的一部分部署和管理的，消除了需要额外的存储筒仓的需求。

使用对象，Nutanix客户可以在现有簇上启用对象存储服务，或设置带有存储密度节点的新群集。

设置

先决条件

一个工作的nutanix对象实例。如果您没有一个，您还在等什么？使用AOS 5.11及以上，每群体免费包含2tib的Nutanix对象。
两个工作的Karbon Kubernetes群集，每个工作人员至少有三名工人。如果您以前从未做过，请看一下这个博客来自我的同事迈克尔·海格（Michael Haigh）。
- 主要的kubernetes群集称为K8S-Prod。
- Kubernetes群集称为K8S-DR。

创建一个水桶

让我们在Nutanix对象中创建一个水桶，Velero将存储备份。

在Prism Central中，单击菜单→服务→对象。
让我们首先创建Velero的访问键。点击访问密钥→增加人数。
选择添加不在目录服务中的人。
您可以使用任何电子邮件地址，我将使用velero@karbon.local。

FCLfQTwPKlpi87bsfHY1yh-wHbiq2A5dCBC-KaG-H7g7JCcd4nNvTYL7vhJg4E5uarWiPObb-y9dvGNQ4JaPNTblcDXH_iv2QeuGYyhiT6_4YEVeUpdp2y-wqJd6o9DvD3g3YJ6P

5.单击下一个和下载键。

6.单击关。

7.单击对象存储然后您的对象实例。我的对象实例名称是Theale。

YBprYtMx0ZSgQuI0nKOGySIz30A3NHoqJdVnN56avQFaRwQ9pk893PoUTrtd4bjsYKej-CzUflKCCh3CPmtDSYZysLR0X89IhCyO0A9R7YrSXzGy5gt-kcG_WDAVmn-dhJqP5LXz

8.单击创建存储桶。

9.例如一个名字，例如velero，然后单击创造。

kBzJWlmqhWrU4oozm0eFVHj_k3gddwxnnpsvuGWO3MFuRgBMsEKmsWhKjfF7DA5s5AzH4VJKxAoHkdhPdaPxeUJVEgTNNbVcjBP_Gu8i3OgMxAQIcIcw9OQlQ72X968--oIPBoQX

10.选择velero桶，然后单击分享。

GZyYDEecQH6hNyKEv85YPxu9ngv7botYZwMmamYl-l8TmimVoe62g_NC6u3q8fRBVruOYIziGmFT4HPSy6s4a8SuZky9vIaGIHIuKTdwgOT49_J4wtBwERbL53QuVl6WF1pb2lQe

11.搜索您之前创建的用户，并添加读/写权限。

IwHFEbD_NgyMSJKa3rcBgy_d7t9hdXOBVwlueWpV7MYIX0B4bTViNT7GoirjtXyObvkvwos9pLpKgHlH_0mMq3UlCw6Yp9B0apwmm9wktH2L8sz0uTobssin4sunDRSaOSpRxFjs

12.单击节省。

Velero准备使用您的水桶！

下载Velero

从您的操作系统下载最新版本的Velero这里。在撰写本博客时，Velero v1.1.0是最新版本。由于我的笔记本电脑是MacBook，我将使用Darwin平台。

安装Velero

1.提取tarball：

tar -xvf <发行射击 - 名称> .tar.gz

2.移至您提取tarball的文件夹：

CD velero- <发行射手名>

3.创建velero特定的凭据文件（凭证-VELERO）在您的本地目录中。您可以从下载的文本文件中检索键：

VI凭证-VELERO

[默认]
aws_access_key_id = 
aws_secret_access_key =

在运行安装命令之前，请确保您有一个工作Kubeconfig对于群集，您要安装Velero（首先是K8S-Prod）。您可以使用kubectl验证这一点。

4.安装AWS云提供商的Velero。Nutanix对象是S3 API兼容，Velero将能够使用S3 API调用将备份存储在其上。使用以下命令（请记住将命令替换为nutanix对象IP地址实例）：

./velero install \
- 提供AWS \
-bucket velero \
- 销售文件./credentials-velero \
-  use-volume-snapshots = false \
-backup-location-config region = us-east-1，s3forcepathstyle =“ true”，s3url = http：//  \ \ \ \
- 使用

5.（可选）检查日志以确保没有错误：

kubectl日志部署/velero -n velero

现在您有一个工作的Velero部署。

6.在我们可以部署应用程序之前，我们需要更新Restic Daemonset，因为库贝特Karbon部署的Kubernetes群集的宿主路径不同（/var/nutanix/var/lib/kubelet）到Restic Standard HostPath（/var/lib/kubelet）。

kubectl -n velero patch daemonset restic -p ='{“ spec”：{“ template”：{“ spec”：{“卷”：[{“ hostpath”：{“ path”：{“ path”：“/var/nutanix/var/var/var/var/var/var/var/var/var/var/var/var/var/var/var/var/var/lib/kubelet/pods“”，“ type”：“”}，“ name”：“ host-pods”}]}}}}}}}}''

7. kubernetes博士群集中的步骤4至6。

是时候部署应用程序并测试备份了。

部署应用程序

该示例是使用状态表，poddisruptionbudgets和podantiaffinity的Zookeeper应用程序。状态群的持续量可以保留Zookeeper分布式数据。对于此示例，您需要一个Karbon Kubernetes群集，其中三名工人。

1. Deploy Zookeeper应用程序：

kubectl应用-f https://raw.githubusercontent.com/kubernetes/website/master/master/content/content/en/examples/application/zookeeper/zookeeper/zookeeper.yaml

2.以下命令说明了如何将注释添加到要备份的吊舱中。

kubectl annotate pod/your_pod_name backup.velero.io/backup-volumes = your_volume_name_name_1，your_volume_name_2，...

就我们的示例而言，由于应用程序是kubernetes状态填充集，因此POD名称是一致的，并且不更改（ZK-0，ZK-1和ZK-2）。该卷相同，名称为datadir对于三个豆荚。

用注释更新三个POD的命令是：

因为我在0 1 2;do kubectl注释pod/zk- $ i backup.velero.io/backup-volumes = datadir;完毕

3.在创建备份之前，让我们在Zookeeper应用程序中创建一些数据。

kubectl exec zk-0 zkcli.sh创建 /karbon岩石

备份应用程序

1.创建velero备份：

./velero备份创建ZK -selector App = ZK

2.检查您的Velero备份状态：

./velero备份描述ZK

如果您在nutanix对象中检查桶，您会发现流量有进出。

6O3EVOUkxt8NcjP505qU7TDRh9zbFhbGIvOX5SRPktrhmhqLDzxjbFlqskymLB-fglSN-y1vyOrKMCLcvUsos58oPMO2kD6Zg7RGuU2gz6xedl6QGrSblao5Tx2OVWxmXEqSXVPc

还原应用程序

让我们恢复备份K8S-DR簇。确保您也在DR集群中安装了Velero。重要的是你有权利Kubeconfig放。

1.获取现有的Velero备份：

./velero备份获取

创建的名称状态有效期存储位置选择器
ZK完成2019-10-19 15:18:47 +0100 BST 29D默认应用= ZK

2.列出kubernetes豆荚默认名称空间以确保您没有运行Zookeeper。如果您有一个，请确保您在Kubernetes博士群中。

kubectl获取豆荚

找不到资源。

3.备份备份：

./velero还原创建-from-backup zk

4.检查还原已成功地观看POD，服务和PVC默认名称空间：

kubectl获取豆荚-W

名称准备状态重新开始年龄
ZK-0 1/1运行0 1M36S
ZK-1 1/1运行0 1M36S
ZK-2 1/1运行0 1M36S

kubectl得到了全部

名称准备状态重新开始年龄
POD/ZK-0 1/1运行0 6M47S
POD/ZK-1 1/1运行0 6M47S
POD/ZK-2 1/1运行0 6M47S

名称类型群集IP外部IP端口年龄
服务/kubernetes clusterip 172.19.0.1  443/tcp 107m
服务/ZK-CS CLUSTERIP 172.19.156.32  2181/TCP 6M47S
服务/ZK-HS群集无 2888/TCP，3888/TCP 6M47S

kubectl获取PVC

名称状态量量访问模式Storageclass年龄
datadir-Zk-0绑定PVC-5350C69E-167F-4A94-839E-98E-98E1FBC1EE38 10GI RWO default-storageclass 7M26S
datadir-zk-1绑定PVC-9AEEB4A5-7E8E-411F-AFE8-B8E4E4E5BD3EEB 10GI RWO RWO default-storageclass 7M26S
Datadir-ZK-2绑定PVC-38717562-2B8B-404A-9BBE-F65C65C65FFF8A29 10GI RWO default-storageclass 7M26S

5.检查我们在生产中创建的数据Kubernetes群集（karbon岩石）在DR集群中可用：

kubectl exec zk-1 zkcli.sh get /karbon

连接到Local主机：2181
2019-10-19 15：09：42,389 [myid：]  - 信息[MAIN：环境@100]  - 客户端环境：Zookeeper.version = 3.4.10-39D3A4F2693333C922ED3DB283BE479BBE479F9DAECAA0F，建立在[...]上[...]
。
。
。
[...]
2019-10-19 15：09：42,503 [myid：]  -  info [main-sendthread（localhost：2181）：clientcnxn $ sendthread@1299]  - 会话在服务器localhost/0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：0：1：2181，sessionID = 0x26DE48881A0E0000，协商超时= 30000

观察者::

观察者状态：综合类型：无路径：null
岩石
CZXID = 0x10000000E
ctime =星期六10月19日14:56:50 UTC 2019
mzxid = 0x10000000e
mtime =星期六10月19日14:56:50 UTC 2019
PZXID = 0x10000000E
cversion = 0
dataversion = 0
aclversion = 0
临时者= 0x0
datalength = 5
numchildren = 0

结论

应该尽可能多地推动备份来使用应用程序内置功能，而不是用于基础架构专有机制。这将确保您保持不可知论，并且无论您在何处运行应用程序，都可以使用相同的过程。

但是，如果您没有其他选择，并且仍然想对整个Kubernetes群集，名称空间或任何特定的Kubernetes对象进行备份，那么Velero是实现此目标的绝佳方法。Nutanix对象用S3 API兼容的存储桶补充Velero，以存储和还原备份。

1个附件

何塞·戈麦斯（Jose Gomez）|技术营销工程师|jose.gomez@nutanix.com |@pipoe2H

Disclaimer<\/strong>: This post is intended for demo purposes and must not be considered production ready. Velero is not included in Nutanix Karbon, hence Nutanix support won\u2019t handle any case related to Velero.<\/p>
\u00a0<\/p>
Overview<\/h2>
One of the main principles of containerised applications is stateless<\/strong>. The reason is not make this application portable and non dependant of any data. In this way you can re-use the same container image on any platform with the same result.<\/p>
\u00a0<\/p>
Because containers started to gain popularity because the portability, scalability, and so on, the community found out a way to containerise stateful applications like databases with the use of local storage, or shared volumes.<\/p>
\u00a0<\/p>
Recover stateless applications is a very straightforward process, you just need to re-apply your manifest file in another cluster. You just need to make sure your manifest file is up-to-date with the latest state running in your cluster. Remember you shouldn\u2019t make changes directly in your cluster, you should update your manifest file with the new changes and apply this into your cluster.<\/p>
\u00a0<\/p>
What about stateful applications? Well, because with containers the data is decoupled from the application, the backup process can be a bit challenging. There are native commands to copy the file system of your persistent volume to another location, or use the native capabilities of your underlying storage to replicate the data to a DR site. But this process requires additional task when you want to restore the data in the DR site, and also depending on what Kubernetes storage plug-in you are using, you may need to export even secret keys to be able to mount the storage in the DR site.<\/p>
\u00a0<\/p>
To streamline the backup and restore process for stateless and stateful applications, Heptio developed Ark, now called Velero. In this blog we will use Velero to backup a ZooKeeper application running in a Nutanix Karbon cluster with Kubernetes. The backup will be pushed to a Nutanix Objects bucket.<\/p>
\u00a0<\/p>
$\"htDRtfSVrVv1TTUgqQE2C7XJ92qtGByhxW17Tgoq_J1jDmkI9s5__cFcwO8E_MSzza8LuU1BUv3GyfOqdb_FnBletrVGgDYGI2sAVutjVu0_cwb2iCdqRaJYneJ5SvpqBkb4RPQg\"$ \u00a0\u00a0<\/p>
What is Velero?<\/h2>
[source: Velero GitHub repo<\/u><\/a>]<\/p>
\u00a0<\/p>
Velero (formerly Heptio Ark) gives you tools to back up and restore your Kubernetes cluster resources and persistent volumes. You can run Velero with a cloud provider or on-premises. Velero lets you:<\/p>
\t
Take backups of your cluster and restore in case of loss.<\/p>\t<\/li>\t
\t
Migrate cluster resources to other clusters.<\/p>\t<\/li>\t
\t
Replicate your production cluster to development and testing clusters.<\/p>\t<\/li><\/ul>
Velero consists of:<\/p>
\t
A server that runs on your cluster<\/p>\t<\/li>\t
\t
A command-line client that runs locally<\/p>\t<\/li><\/ul>
Velero has support for backing up and restoring Kubernetes volumes using a free open-source backup tool called restic. This support is considered beta quality. Please see the list of limitations<\/u><\/a> to understand if it currently fits your use case.<\/p>
\u00a0<\/p>
What is restic?<\/h2>
[source: restic GitHub repo<\/u><\/a>]<\/p>
\u00a0<\/p>
restic is a backup program that is fast, efficient and secure. It supports the three major operating systems (Linux, macOS, Windows) and a few smaller ones (FreeBSD, OpenBSD).<\/p>
\u00a0<\/h2>
What is Nutanix Objects?<\/h2>
Nutanix Objects $\"\u2122\ufe0f\"$ is a software-defined object storage solution that non-disruptively scales-out while lowering overall costs. It\u2019s designed with an S3-compatible REST API interface to handle terabytes to petabytes of unstructured data, all from a single namespace. Objects is designed for backup, long term retention\/archiving, and cross-region devops teams. It\u2019s deployed and managed as part of the Nutanix Enterprise Cloud Platform, eliminating the need for additional storage silos.<\/p>
\u00a0<\/p>
With Objects, Nutanix customers can enable object storage services on existing clusters or set up new clusters with storage dense nodes.<\/p>
\u00a0<\/p>
Setup<\/h2>
Prerequisites<\/h3>
\t
A working Nutanix Objects instance. If you don\u2019t have one, what are you waiting for? With AOS 5.11 and above, 2TiB of Nutanix Objects is included for free on a per-cluster basis.<\/p>\t<\/li>\t
\t
Two working Karbon Kubernetes clusters with at least three workers each. If you haven\u2019t done it before, have a look at this blog<\/u><\/a> from my colleague Michael Haigh.<\/p>\t
\t\t
Primary Kubernetes cluster called K8s-Prod<\/strong>.<\/p>\t\t<\/li>\t\t
\t\t
DR Kubernetes cluster called K8s-DR<\/strong>.<\/p>\t\t<\/li>\t<\/ul><\/li><\/ul>
Create a Bucket<\/h3>
Let\u2019s create a bucket in Nutanix Objects where Velero will store the backups.<\/p>
\t
In Prism Central, click Menu<\/strong> \u2192 Services<\/strong> \u2192 Objects<\/strong>.<\/p>\t<\/li>\t
\t
Let\u2019s create first the access keys for Velero. Click Access Keys<\/strong> \u2192 Add People<\/strong>.<\/p>\t<\/li>\t
\t
Choose Add people not in a directory service<\/strong>.<\/p>\t<\/li>\t
\t
You can use any email addresses, I\u2019ll use velero@karbon.local<\/strong>.<\/p>\t<\/li><\/ol>
$\"FCLfQTwPKlpi87bsfHY1yh-wHbiq2A5dCBC-KaG-H7g7JCcd4nNvTYL7vhJg4E5uarWiPObb-y9dvGNQ4JaPNTblcDXH_iv2QeuGYyhiT6_4YEVeUpdp2y-wqJd6o9DvD3g3YJ6P\"$ <\/p>
5.Click Next<\/strong> and Download Keys<\/strong>.<\/p>
6.Click Close<\/strong>.<\/p>
7.Click Object Stores<\/strong> and later your Objects instance. My Object instance name is Theale<\/em>.<\/p>
$\"YBprYtMx0ZSgQuI0nKOGySIz30A3NHoqJdVnN56avQFaRwQ9pk893PoUTrtd4bjsYKej-CzUflKCCh3CPmtDSYZysLR0X89IhCyO0A9R7YrSXzGy5gt-kcG_WDAVmn-dhJqP5LXz\"$ <\/p>
8.Click Create Bucket<\/strong>.<\/p>
9.Give a name, for example velero,<\/strong> and click Create<\/strong>.<\/p>
$\"kBzJWlmqhWrU4oozm0eFVHj_k3gddwxnnpsvuGWO3MFuRgBMsEKmsWhKjfF7DA5s5AzH4VJKxAoHkdhPdaPxeUJVEgTNNbVcjBP_Gu8i3OgMxAQIcIcw9OQlQ72X968--oIPBoQX\"$ <\/p>
\u00a0<\/p>
10.Select the velero<\/em> bucket, and click Share<\/strong>.<\/p>
$\"GZyYDEecQH6hNyKEv85YPxu9ngv7botYZwMmamYl-l8TmimVoe62g_NC6u3q8fRBVruOYIziGmFT4HPSy6s4a8SuZky9vIaGIHIuKTdwgOT49_J4wtBwERbL53QuVl6WF1pb2lQe\"$ <\/p>
11.Search for the user you created before, and add Read<\/strong>\/Write<\/strong> permissions.<\/p>
$\"IwHFEbD_NgyMSJKa3rcBgy_d7t9hdXOBVwlueWpV7MYIX0B4bTViNT7GoirjtXyObvkvwos9pLpKgHlH_0mMq3UlCw6Yp9B0apwmm9wktH2L8sz0uTobssin4sunDRSaOSpRxFjs\"$ <\/p>
12.Click Save<\/strong>.<\/p>
Your bucket is ready to be used by Velero!<\/p>
\u00a0<\/p>
Download Velero<\/h3>Download the latest version of Velero for your operating system from here<\/u><\/a>. At the time of writing this blog Velero v1.1.0 is the latest release. I\u2019ll be working with the Darwin platform since my laptop is a MacBook.<\/p>
\u00a0<\/p>
Install Velero<\/h3>
1.Extract the tarball:<\/p>
tar -xvf <RELEASE-TARBALL-NAME>.tar.gz<\/code><\/pre>2.Move to the folder where you have extracted the tarball:<\/p>
cd velero-<RELEASE-TARBALL-NAME><\/code><\/pre>3.Create a Velero-specific credentials file (credentials-velero<\/strong>) in your local directory. You can retrieve the keys from the downloaded text file:<\/p>
vi credentials-velero<\/code><\/pre>
[default] aws_access_key_id = <OBJECTS-ACCESS-KEY> aws_secret_access_key = <OBJECTS-SECRET-KEY><\/code><\/pre>Before you run the install command, make sure you have a working <\/u>KUBECONFIG<\/u><\/strong> for the cluster you want to install Velero (K8s-Prod first). You can use <\/u>kubectl<\/em><\/u> to verify this.<\/u><\/p> 4.Install Velero for the AWS cloud provider. Nutanix Objects is S3 API compatible, Velero will be able to store the backups on it using S3 API calls. Use the following command (remember to replace the command with your Nutanix Objects IP address instance):<\/p>
.\/velero install \\ --provider aws \\ --bucket velero \\ --secret-file .\/credentials-velero \\ --use-volume-snapshots=false \\ --backup-location-config region=us-east-1,s3ForcePathStyle=\"true\",s3Url=http:\/\/<OBJECTS-INSTACE-IP-ADDRESS> \\ --use-restic<\/code><\/pre>\u00a05.(Optional) Check the logs to make sure there are no errors:<\/p>kubectl logs deployment\/velero -n velero<\/code><\/pre>Now you have a working Velero deployment.<\/p> 6.Before we can deploy an application, we need to update Restic DaemonSet because the kubelet<\/em> hostPath for Kubernetes clusters deployed by Karbon is different (\/var\/nutanix\/var\/lib\/kubelet<\/strong>) to the Restic standard hostPath (\/var\/lib\/kubelet<\/strong>).<\/p>
kubectl -n velero patch daemonset restic -p='{\"spec\":{\"template\":{\"spec\":{\"volumes\":[{\"hostPath\":{\"path\":\"\/var\/nutanix\/var\/lib\/kubelet\/pods\",\"type\":\"//www.jhbzcj.com/next/karbon-kubernetes-orchestration-30/\"},\"name\":\"host-pods\"}]}}}}'<\/code><\/pre>\u00a07.Repeat steps 4 to 6 in your Kubernetes DR cluster.<\/p> It\u2019s time to deploy an application and test a backup.<\/p> \u00a0<\/p>
Deploy Application<\/h3>
The example is a ZooKeeper application using StatefulSets, PodDisruptionBudgets, and PodAntiAffinity. The persistent volumes for the StatefulSets keeps the ZooKeeper distributed data. For this example you need a Karbon Kubernetes cluster with three workers<\/u>.<\/p>
1.Deploy the ZooKeeper app:<\/p>
kubectl apply -f https:\/\/raw.githubusercontent.com\/kubernetes\/website\/master\/content\/en\/examples\/application\/zookeeper\/zookeeper.yaml<\/code><\/pre>\u00a02.The following command explains how to add annotations to the pods you want to backup.<\/p>
kubectl annotate pod\/YOUR_POD_NAME backup.velero.io\/backup-volumes=YOUR_VOLUME_NAME_1,YOUR_VOLUME_NAME_2,...<\/code><\/pre>For our example, because the application is a Kubernetes StatefulSet, the pod names are consistent and don\u2019t change (zk-0, zk-1 and zk-2). The same for the volume, the name is datadir<\/strong> for the three pods.\u00a0<\/p> The command to update the three pods with the annotations is:<\/p>
for i in 0 1 2; do kubectl annotate pod\/zk-$i backup.velero.io\/backup-volumes=datadir; done<\/code><\/pre>\u00a03.Before creating a backup, let\u2019s create some data in our ZooKeeper application.<\/p>
kubectl exec zk-0 zkCli.sh create \/karbon rocks<\/code><\/pre>\u00a0<\/h3>
Backup Application<\/h3>
1.Create a Velero backup:\u00a0<\/p>
.\/velero backup create zk --selector app=zk<\/code><\/pre>\u00a02.Check your Velero backup status:<\/p>
.\/velero backup describe zk<\/code><\/pre>If you check the bucket in Nutanix Objects, you will see there has been traffic coming in and out.<\/p> <\/p> \u00a0<\/h3>
Restore Application<\/h3>
Let\u2019s restore the backup in the K8s-DR<\/strong> cluster. Make sure you have installed Velero in your DR cluster too. It\u2019s important you have the right KUBECONFIG<\/strong> set.<\/p>
1.Get the existing Velero backups:<\/p>
.\/velero backup get<\/code><\/pre>NAME STATUS CREATED EXPIRES STORAGE LOCATION SELECTOR zk Completed 2019-10-19 15:18:47 +0100 BST 29d default app=zk <\/code><\/pre>\u00a02.List the Kubernetes pods in the default<\/strong> namespace to make sure you don\u2019t have ZooKeeper running. If you have one, make sure you are on the DR Kubernetes cluster.<\/p> kubectl get pods<\/code><\/pre>No resources found.<\/code><\/pre>\u00a03.Restore the backup:<\/p> .\/velero restore create --from-backup zk<\/code><\/pre>\u00a04.Check the restore has been successful watching the pods, services and PVCs in the default<\/strong> namespace:<\/p> kubectl get pods -w<\/code><\/pre>NAME READY STATUS RESTARTS AGE zk-0 1\/1 Running 0 1m36s zk-1 1\/1 Running 0 1m36s zk-2 1\/1 Running 0 1m36s <\/code><\/pre>kubectl get all<\/code><\/pre>NAME READY STATUS RESTARTS AGE pod\/zk-0 1\/1 Running 0 6m47s pod\/zk-1 1\/1 Running 0 6m47s pod\/zk-2 1\/1 Running 0 6m47s NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE service\/kubernetes ClusterIP 172.19.0.1 <none> 443\/TCP 107m service\/zk-cs ClusterIP 172.19.156.32 <none> 2181\/TCP 6m47s service\/zk-hs ClusterIP None <none> 2888\/TCP,3888\/TCP 6m47s <\/code><\/pre>kubectl get pvc<\/code><\/pre>NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE datadir-zk-0 Bound pvc-5350c69e-167f-4a94-839e-98e1fbc1ee38 10Gi RWO default-storageclass 7m26s datadir-zk-1 Bound pvc-9aeeb4a5-7e8e-411f-afe8-b8e4e5bd3eeb 10Gi RWO default-storageclass 7m26s datadir-zk-2 Bound pvc-38717562-2b8b-404a-9bbe-f65c65ff8a29 10Gi RWO default-storageclass 7m26s<\/code><\/pre>\u00a05.Check if the data we created in our production Kubernetes cluster (karbon rocks<\/strong>) is available in the DR cluster:<\/p> kubectl exec zk-1 zkCli.sh get \/karbon<\/code><\/pre>Connecting to localhost:2181 2019-10-19 15:09:42,389 [myid:] - INFO [main:Environment@100] - Client environment:zookeeper.version=3.4.10-39d3a4f269333c922ed3db283be479f9deacaa0f, built on [...] . . . [...] 2019-10-19 15:09:42,503 [myid:] - INFO [main-SendThread(localhost:2181):ClientCnxn$SendThread@1299] - Session establishment complete on server localhost\/0:0:0:0:0:0:0:1:2181, sessionid = 0x26de4881a0e0000, negotiated timeout = 30000 WATCHER:: WatchedEvent state:SyncConnected type:None path:null rocks cZxid = 0x10000000e ctime = Sat Oct 19 14:56:50 UTC 2019 mZxid = 0x10000000e mtime = Sat Oct 19 14:56:50 UTC 2019 pZxid = 0x10000000e cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 5 numChildren = 0<\/code><\/pre>\u00a0<\/h2> Conclusion<\/h2>Backups should be pushed as much as possible to use the application built-in capabilities for this than to infrastructure proprietary mechanisms. This will ensure you keep agnostic and the same process will work regardless of where you run your application.<\/p> \u00a0<\/p> But if you don\u2019t have another choice and still want to take a backup of your entire Kubernetes cluster, namespace or any specific Kubernetes objects, Velero is a very good way to achieve this. Nutanix Objects complements Velero with a S3 API compatible bucket to store and restore backups.\u00a0<\/p> \u00a0<\/p> \u00a0<\/p>","quoteUsername":"JoseNutanix","translations":{"Common":{"like":"Like","unlike":"Unlike"},"Forum":{"Quote":"Quote","Share":"Share"}}}"> 喜欢引用分享

该主题已关闭以供评论

由内部提供动力条款和条件注册已经有一个帐户？登录使用您的帐户登录登录社区使用您的帐户登录输入您的用户名或电子邮件地址。我们将向您发送带有指令的电子邮件以重置您的密码。用户名或电子邮件返回概述扫描病毒文件。抱歉，我们仍在检查该文件的内容，以确保它可以安全下载。请在几分钟后再试一次。好的该文件无法下载抱歉，我们的病毒扫描仪检测到该文件无法安全下载。好的 Learn more about our cookies.<\/a>","cookiepolicy.button":"Accept cookies","cookiepolicy.button.deny":"Deny all","cookiepolicy.link":"Cookie settings","cookiepolicy.modal.title":"Cookie settings","cookiepolicy.modal.content":"We use 3 different kinds of cookies. You can choose which cookies you want to accept. We need basic cookies to make this site work, therefore these are the minimum you can select. Learn more about our cookies.<\/a>","cookiepolicy.modal.level1":"Basic Functional","cookiepolicy.modal.level2":"Normal Functional + analytics","cookiepolicy.modal.level3":"Complete Functional + analytics + social media + embedded videos"}}}">

概述

什么是Velero？

什么是restic？

什么是nutanix对象？

设置

先决条件

创建一个水桶

下载Velero

安装Velero

部署应用程序

备份应用程序

还原应用程序

结论

1个附件

Setup<\/h2>

\u00a0<\/h3>

Backup Application<\/h3>1.Create a Velero backup:\u00a0<\/p>

\u00a0<\/h3>

\u00a0<\/h2>

注册

登录社区

扫描病毒文件。

该文件无法下载