DevOps Tutorials | Kubernetes cluster backup and restore with Velero - Kubernetes cluster backup

Summary

This video discusses Valero, a backup and restore tool for Kubernetes clusters. It emphasizes the importance of backups for disaster recovery and data preservation, regardless of a cluster's resilience. The presenter introduces Valero, which utilizes cloud providers' block storage snapshot features to snapshot persistent volumes and restore cluster objects. The video covers backup strategies, including full backups and namespace-specific backups, and demonstrates how to install Valero, take backups, and restore from backups. It also shows a practical scenario where a namespace is deleted and then recovered using Valero.

Takeaways

• 😀 Kubernetes backup is crucial for disaster recovery and maintaining data integrity.
• 🛠 Valero is a backup tool for Kubernetes clusters that leverages cloud providers' block storage snapshot features.
• 📈 Valero supports full backups, namespace-specific backups, and resource type backups, offering flexibility in backup strategies.
• ⏰ Backups can be scheduled to run periodically, ensuring ongoing data protection.
• 🔍 Valero is useful for recovering from disasters, such as accidental namespace deletions or failed Kubernetes API upgrades.
• 🔄 It facilitates environment replication and major updates, such as moving from development to production.
• 💾 Backups include both Kubernetes state (etcd) and application data on persistent volumes.
• 📋 Prerequisites for using Valero include a running Kubernetes cluster and the installation of the Valero tool.
• 🔧 The Valero client and server must be installed, and the server should be running to perform backup and restore operations.
• 🔄 The process of restoring a deleted namespace demonstrates Valero's ability to recover from data loss scenarios.

Q & A

What is the importance of backup in Kubernetes?

Backup in Kubernetes is crucial for disaster recovery and ensuring data integrity. It allows you to restore your cluster objects and persistent volumes to a previous state in case of failures, data loss, or accidental deletions.

Why is it essential to have a backup and disaster recovery plan for Kubernetes clusters?

A backup and disaster recovery plan are essential to recover from catastrophic events, such as accidental deletions, cluster failures, or network issues, and to maintain the availability and resilience of your applications.

What is Valero and how does it relate to Kubernetes backup?

Valero is a backup tool for Kubernetes clusters that takes snapshots of your clusters' persistent volumes using cloud providers' block storage snapshot features. It allows you to perform full backups or backups of specific namespaces or resource types.

Can Valero be used for periodic backups in Kubernetes?

Yes, Valero can be used to schedule backups to be executed periodically, providing a flexible approach to backup strategies depending on the needs of the environment.

What are some scenarios where backup and recovery in Kubernetes are necessary?

Backup and recovery in Kubernetes are necessary in scenarios such as accidental namespace deletion, failed Kubernetes API upgrades, network outages, or when replicating environments from development to production.

What are the prerequisites for using Valero to backup a Kubernetes cluster?

The prerequisites for using Valero include having a running Kubernetes cluster and installing the Valero tool in your environment.

How can one verify if a backup taken by Valero has been completed successfully?

To verify if a Valero backup has been completed, you can use the command `velero backup describe <backup-name>` which will show the details including the completion status, start time, and expiry date of the backup.

What happens if a namespace is accidentally deleted in a Kubernetes cluster?

If a namespace is accidentally deleted, it can lead to loss of services and deployments within that namespace. However, with Valero, you can restore the deleted namespace and its resources from a previously taken backup.

How does Valero help in recovering from a disaster in a Kubernetes cluster?

Valero helps in disaster recovery by allowing you to restore your cluster to a previous state using the snapshots taken during the backup process. This can be done by issuing a restore command with the specific backup name.

What are the steps to restore a Kubernetes cluster using Valero after a disaster?

To restore a Kubernetes cluster using Valero, you first ensure Valero is installed and running, then use the command `velero restore create --from-backup <backup-name>` to initiate the restore process, and finally verify the restoration by checking the restored resources.

Outlines

00:00

🔒 Importance of Backup in Kubernetes

The speaker introduces the topic of backup and restore tools in Kubernetes, emphasizing the necessity of backing up data before going to production. They highlight the importance of having a solid backup and disaster recovery plan, regardless of the cluster's resilience. The tool 'Velero' is introduced as a solution for backing up Kubernetes clusters, which takes snapshots of persistent volumes using cloud providers' block storage snapshot features. The speaker also discusses the different types of backups one can perform with Velero, such as full backups or backups of specific namespaces or resource types, and the flexibility to schedule backups.

05:02

🛠 Setting Up and Using Velero

The speaker provides a step-by-step guide on how to set up and use Velero. They begin by pulling down the Velero GitHub repository and installing the Velero client. The process involves downloading the Velero server, dealing with errors that arise during installation, and ensuring that the prerequisites for backup are met, such as having a running cluster and Velero installed. The speaker also mentions the documentation available on the Velero website for further guidance on installation and backup procedures.

10:03

🚀 Deploying an Application and Taking a Backup

After successfully installing Velero, the speaker proceeds to deploy an Nginx application within the Kubernetes cluster. They demonstrate how to check the deployment status and ensure that Velero and the Nginx application are both deployed correctly. The speaker then walks through the process of taking a backup using the 'velero backup create' command, explaining how to verify the completion and details of the backup.

15:04

🔄 Simulating Disaster and Restoring from Backup

The speaker simulates a disaster scenario by deleting a namespace that contains the Nginx application. They demonstrate how to delete the namespace using 'kubectl delete namespace' and verify that the services and deployments are gone. After the deletion, the speaker shows how to restore the deleted namespace and its resources using Velero's 'restore create' command. They verify the restoration by checking the namespace and services, confirming that the cluster has been successfully restored from the backup.

20:08

🏁 Conclusion

The final paragraph is a brief conclusion or a placeholder indicating the end of the discussion. It does not contain any content, but it signifies the completion of the video script.

Keywords

💡Valero

Valero is a backup and restore tool used in Kubernetes environments. It is crucial for disaster recovery and data preservation, as it allows users to take snapshots of cluster states and persistent volumes. In the video, Valero is introduced as a tool for creating backups of Kubernetes clusters, which can be full or partial, and can be scheduled for periodic execution. The script mentions using Valero to take backups before going to production, highlighting its importance in ensuring data integrity and cluster recovery.

💡Kubernetes

Kubernetes is an open-source container orchestration platform for automating deployment, scaling, and management of containerized applications. It is the central theme of the video, as the script discusses the importance of backing up Kubernetes clusters to ensure high availability and resilience. The video script emphasizes the need for a solid backup and disaster recovery plan regardless of how robust the Kubernetes cluster is.

💡Backup

Backup refers to the process of creating a copy of data to prevent data loss in case of system failures or disasters. In the context of the video, backup is a critical operation for Kubernetes clusters to ensure that data and cluster states can be restored to a previous state. The script discusses various scenarios where backups are essential, such as before going to production or when migrating clusters.

💡Disaster Recovery

Disaster Recovery (DR) is a strategy to prepare for potential disasters, such as data loss or system failures, by having a plan to recover and restore critical functions and data. The video script mentions the importance of a disaster recovery plan in conjunction with backups, emphasizing that even highly available Kubernetes clusters need a strategy to recover from catastrophic events.

💡Persistent Volume

Persistent Volume (PV) in Kubernetes is a piece of storage in the cluster that has been provisioned by an administrator or dynamically provisioned using storage classes. The script explains that Valero takes snapshots of persistent volumes, which are crucial for backup because they store the stateful data of applications running in the cluster.

💡Namespace

A Namespace in Kubernetes is a logical partition that allows resources to be grouped together. The video script uses the term 'namespace' when discussing the backup and restore process, highlighting that backups can be taken at the namespace level, allowing for more granular control over what is backed up and restored.

💡Snapshot

A snapshot, in the context of the video, refers to a point-in-time capture of a disk's data. Valero uses cloud provider block storage snapshot features to create snapshots of Kubernetes clusters' persistent volumes, which can later be used to restore the cluster to a previous state. The script mentions snapshots as a core feature of the backup process.

💡Cluster Migration

Cluster Migration is the process of moving a Kubernetes cluster from one environment to another, such as from a development environment to a production environment. The video script discusses the use of backups during cluster migrations to ensure that the new environment has the same state as the original.

💡Deployment

In Kubernetes, a Deployment provides declarative updates to applications and their resources. The video script refers to deployments when demonstrating the backup and restore process, showing how Valero can back up and restore deployments within a namespace.

💡Data Integrity

Data Integrity ensures that data is accurate, consistent, and remains unaltered. The video script underscores the importance of maintaining data integrity through backups, especially before critical operations like going to production or performing major updates. It is a fundamental aspect of ensuring that the backup and restore process is reliable.

Highlights

Valero is a backup and restore tool used in Kubernetes environments.

Backup is crucial for data and cluster security in production environments.

Regardless of cluster resilience, a solid backup and disaster recovery plan are essential.

Valero takes snapshots of Kubernetes clusters' persistent volumes using cloud provider features.

Valero supports full backups, namespace-specific backups, and scheduled backups.

Backups are necessary for disaster recovery, accidental deletions, and API upgrade failures.

Backups are also important for environment replication and major updates.

Backups should include Kubernetes state, persistent volumes, and application data.

Prerequisites for backup include an up-and-running cluster and the installation of Valero.

Valero's GitHub repository provides installation instructions and use cases.

The video demonstrates how to download, install, and verify Valero.

The process of deploying an Nginx application in Kubernetes is shown.

A step-by-step guide on how to take a backup of Kubernetes resources using Valero.

Demonstration of deleting a namespace to simulate a disaster scenario.

Restoring a deleted namespace using Valero to showcase the recovery process.

Verification of successful restoration by checking the restored namespace and services.

Transcripts

in this video I'm going to talk about

the Valero one of the backup

and restore tools which we are using in

kubernetes so um uh we know

um whenever you work in any environments

we always talk about the backup right we

always take to whether we have taken the

backup of our data or not so

so uh whenever you go to the Productions

we always take the backup of your data

or the cluster so here you know in this

kubernetes backup are also important

okay so it's very important to take the

back up before going to the Productions

or any any other environments data is

very important so uh it's not uh it's no

matter how highly available or resilient

your kubernetes cluster is it is

important to have a solid backup and you

store plan and just Disaster Recovery

plan okay so uh that's why I uh going to

talk about here one of the tool okay

that's called uh

Bolero for backup in the store for our

applications okay so uh we are going to

uh The Tool uh let's uh talk about what

about the tool is doing okay so Valero

has become one of the backup tools for

kubernetes cluster nowadays because uh

this is one of for the cnca project and

uh it's a takes the snapshot of your

clusters

persistent volume uh using one of your

uh Cloud providers block storage

snapshot features and uh later you can

restore your cluster objects and

persistent volumes to a previous state

basically and so so uh you can use

Valero to perform full backups full

backups or you can backups of some of

the name space age or resource types or

can be scheduled backup to execute them

periodically so it depends how means or

what kind of backup you want to take

whether the name space or you want to

critically take so it's all depends on

up to you okay but why we want to take

that back up the Clusters Okay so

backup you know right so where why you

need a backup and Recovery mechanism

cluster let's suppose something goes

wrong so to recover from the disaster or

someone accidentally uh let's suppose

delete it the name space or your

kubernetes API upgrade failed and you

have to revert back your network went

down now your cluster is not reachable

um

or any Mutual disasters happen right so

that's why to recover from the that

states the backup is important okay

also when you replicate the environments

uh from development to staging or

um whenever you doing the major update

upgrade from um tape to acceptance or

Productions you take the backup position

or when you are doing the migration of

kubernetes clusters from one environment

to another environments Okay so what are

the things you take the backup here

okay so when you talk about the backup

basically uh

so when you talk about the backup two

things you need to backup kubernetes

stores uh in a state in its City so uh

basically it just stores the master its

City and relevant certificate which we

should be backed up okay and then the

application data okay it's a

also

yeah

yeah it's also uh take the backup of

your application data persistent volume

uh because in let's suppose you will be

having the any application running your

clusters and uh uh if your cluster goes

down or it's not reachable it's very

important to have the backup people

applications right so

um yeah so

what is the prerequisites uh to have the

backup your cluster should be up and

running okay and uh you have to install

uh the Valero in your environment so so

uh you can go to the Valero site here

okay Valero dot IO and you can see the

use cases here Disaster Recovery cluster

migrations backup reference so these are

the use cases this have also explained

about how you can do the installations

how you can took the backup everything

is has been um explained here in this

page so you can go and uh

uh read about this document so uh this

uh web page uh let's go uh to the demo

uh how we can take the backup how you

restores we will also see one scenario

wherever I will uh I will uh delete some

namespace or cluster and how I will

recover that so so uh let's uh start

yeah so uh I'm going to start the lab

here so first what I will do I will uh

uh pull down the Valero GitHub repo okay

um to start so I have also prepared uh

the document here uh how to take backup

migrate our uh

kubernetes clusters I have prepared here

okay so I will basically follow the

steps I have written down the all the

commands here

I have also written here so let's uh

first uh

clone the

apto Valero

okay so what we'll do uh it will pull

down the value of GitHub repo okay uh

yeah once it is done

um you what you have to do you have to

download and install the uh your client

here okay so

you can

download and install the Valero client

from here

okay

so uh once it's done uh we will check

the version of our value which

installed in this

cluster okay so let's uh

first download it

once it is downloaded

then

we will extract this top file here

meanwhile I will go and copy this

command

yeah now it will take some time because

it's a very big file

uh you can

go and refer this is the uh

Valero GitHub project okay so uh you saw

right I I

downloaded or pulled this GitHub project

github.com here

so this is the project which I uh

downloaded uh in my local in my cluster

basically okay so

um it is downloaded okay so now you

extract it okay

so once it is done I will add Valor

directory to uh the path okay

so let's wait for some time and

let's go back

and

copy this

yeah so it is done

extracted

Okay so

you can see we are getting some error

here okay uh because why I'm getting

this error uh because uh this Valero is

not started yet okay

um so we will start that we will create

the Valerio specific credential file

here first

okay so let's do it here

it's done then we will start the server

here

we will uh get all the steps on the

Valero web page okay

and

let's see again I'm getting the error

let's see how can I uh

solve this one okay it's something I

have missed okay so I have not installed

the fellow A8 so let's install it now

I got I have just installed the client

now I am installing the server

so uh once it is done you can see this

uh deployment Valero created Valero is

installed so when you now run the Valero

version

it should now keep

the W details about your version So

currently we are getting version one one

zero so uh uh now you will not get any

error because your Valero is installed

you can see here showing this Valero is

installed

okay so what next to what I'm going to

do uh once uh this Valero is installed

um say our main purpose uh okay again

I'm getting some error why waiting for

server status request to process

wait

uh sometimes it takes time yeah so I

think it will show it should work

uh time weight okay so what I will do I

will uh again install this uh sometimes

uh

um we Face ratio

so let's see

yes so you see uh now I got the client

and server detail so don't worry uh when

you get uh this type of error uh again

you install uh the Bell Arrow okay so

we'll

um because I I faced this issue before

so I know um how to solve this so you

can do the similar things here now what

I'm going to do here now I'm going to

uh deploy here one nginx application uh

in my cluster

yeah so if you go and see uh Cube CTL

get

deployments you see there is no resource

so what I am going to do I am going to

deploy one um

application nginx applications

why it's not

Okay so

deployments

why it's not showing

keeps it will get the comments

where is it installed in space okay so

basically it's uh deployed in nginx

hyphen

um example name space so what I will do

I will go to

and in x

okay so you see uh

there is even application named nginx

iPhone deployment is deployed here you

can also uh see what are the namespace

is available here so Cube CTL get NS so

this is the one which uh uh and the next

application is deployed in this name

space and this is the Valero or name

space so where your Valero is installed

okay so what I'm going to do next uh we

already checked that your Valero and

nginx is deployed okay

um you can

also check using this command here

so

you see value is deployed also okay what

I am going to next uh we also checked

our nginx is deployed

um

here we already checked this now we will

uh take the uh backup okay so uh once uh

Valero and nginx deployment is

successfully created what I'm going to

backup our resources okay so to take the

backup of our resources this is a

command where a row backup create nginx

iPhone backup so let's go

okay so what you see here it's saying

there's Valero request nginx backup

submitted successfully and uh you can

describe this command uh to see more

about uh uh what

is backed up here okay so you can see

yeah yeah with this command you can

verify if the backup has been completed

or not so you can see it's showing that

it's it's completed so your backup is

completed uh it will show you the date

and time or what time it will um started

what are the expiration day expiry date

so all these things uh you can see here

okay so what I will do now

um we have taken the backup everything

now we will delete the namespace okay so

that's our purpose right we want to

check whether we are able to uh recover

our

like uh

cluster or not right so what one in this

example I am going to uh do some

disasters like I'm going to uh delete

one name space so you know right we have

um namespace

uh called nginx hyphen example where our

nginx application is deployed right so

what I'm going to do so we will going to

delete now this so the command is Cube

CTL

um delete command here

you can also type

um if you are so fast Cube CTL

uh Delete

namespace

and then your name is this name

so what's our name space name name space

name is nginx hyphen example let's do

copy and paste here

Okay so

it will get deleted in some time

uh okay so once it is deleted we will uh

check that and the next service and

deployment are gone right so

okay

so how will you check that so it's very

simple again Cube CTL get deployment and

this name space right

so let's verify that

it's saying that sorry it's so my

mistake I have in the wrong command here

okay so no this was found because uh why

because I have already deleted right so

what I will do now I have to recover I

have already taken the backup of this so

I I can um easily um restore this right

so you will also give verify the

services so everything is deleted you

you go up obviously we have deleted the

uh name space so we'll not find anything

in this particular name space so what

I'm going to do I will uh restore

uh the

cluster okay so how will you do this uh

uh command first we also check this view

uh verify everything

yes nothing is there

okay so nothing is there

so this is the command to restore so I'm

going to uh uh restore here

so how will you restore the command is

Valero restore create hyphen iPhone from

backup and the backup name this is the

name of the backup right so what you see

here uh restore request submitted

successfully uh run Valero restore

describe all these things so let's do uh

we will see

verify that it's a restore or not first

okay

so it's saying completed so now next uh

run the command

restore get

okay

okay so you can see it's uh restored

completed zero warning zero errors and

now we restore so we can go and check

whether our name space is back or not

okay so uh I will quickly go and then

qctl

get NS

so you see name nginx example is back

here okay

yes so this is the service okay this

service is available here

now

Cube CTL get already we got this name

space yeah so we verify that we restored

our cluster here okay uh we just wrote

the name space here so it's it's very

simple ah