[ Kube 19 ] Performing Rolling Updates in Kubernetes

In this video I will show you how you can update your deployments in your Kubernetes cluster using Rolling Updates concept. I have used microk8s for this demonstration.
Github: github.com/justmeandopensourc...
Thanks for watching my video. For any issues/questions/feedback, please drop me a comment and don't forget to subscribe to my channel.
If you wish to support me:
www.paypal.com/cgi-bin/webscr...
Thanks again.

[@FreshersKosam]

Super Bro, I got a task in my office to check out the roll out strategy. this absolutely covers 200% requirement

[@justmeandopensource]

Hi Siva, thanks for watching. Cheers.

[@rohanrana5579]

Loved the different cases and scenarios Thanks venkat, keep up the good work.

[@justmeandopensource]

Hi Rohan, Thanks for watching.

[@mini-sunnykifoodpanti1371]

wow super stuff! One shop stop for all concepts. It would have taken me a whole day to find & understand all this from official documentation. Thanks a lot Venkat :)

[@justmeandopensource]

You are welcome. Thanks for watching.

[@mathewkargarzadeh3158]

Thanks. Very Good illustration between Recreate and Rolling update. Thank you again !!!

[@justmeandopensource]

You are welcome and thanks for watching this video. Cheers.

[@janl.6568]

Love your videos !!

[@justmeandopensource]

Hi Jan, thanks for watching. Cheers.

[@kalpsoni1873]

Hey Dude.....Really superb video....Talks about all day to day challenges in production...Keep up the good work ...

[@justmeandopensource]

Hi Kalp, thanks for watching.

[@KloudEnrich_RameshkumarB]

Excellent Explanation bro

[@justmeandopensource]

Hi Ram, thanks for watching.

[@md.zakirhossain6210]

Your videos are awesome. Thanks for your efforts.

[@justmeandopensource]

Hi Zakir, thanks for watching. Cheers.

[@AshokReddy-er5gl]

You explained very well

[@justmeandopensource]

Thanks. Hope it helped.

[@devanshuoza8561]

Very helpful video. Thank you

[@justmeandopensource]

Hi Devanshu, thanks for watching this video.

[@mohammedsuhailbasha4860]

Great explaination bro. Thanks lot. Please make and upload videos on Kubernetes calico networking deep dive, daemonset real world scenario and Kubernetes network policies.

[@justmeandopensource]

Hi Mohammed, thanks for watching. I will add it to my list. Cheers.

[@romantsyupryk3009]

Thanks so much for this tutorial.

[@justmeandopensource]

Hi Roman, thanks for watching.

[@Praveen-xx7iy]

Hi Venkat, I couldn't able to find video on blue green deployment. Previously I could able to see it.

[@justmeandopensource]

Hi Praveen, thanks for watching. Blue green deployment in my playlist? I don't think I did that video.

[@bashardlaleh2110]

what strategy does rollout restart follow and is it good for production? thanks for the high-quality content.

[@irfanansari6448]

Hi Venkat, Thanks for this video but I was thinking what if we have pods with multiple containers using different images in deployment. How we are going to achieve it.?

[@zhihaowang5410]

Hi thanks for the video, I like it, I have one question about the rolling updates, For example, I have a deployment app, but the deployment app only have one pod, I do not use replicaSet for the deployment, then I want to live upgrade the pod image, then how can I do the rolling update? is that the same step as the video?

[@zaheerhussain5311]

Hi Do you have any video on auto rollback in kubernetes when deployment becomes fails? Regards Zaheer

[@CrazySkillz15]

Great explanantion! Thanks for posting this in! I had a question. What shell commands did you use to get a dashboard of k8s services deployments and pods like that such that it updates every 2 seconds? I know --watch can be used but its not very informative for my usecase.

[@justmeandopensource]

Hi Abhijeet, thanks for watching this video. I used the good old watch command for that. Any linux command prefixed with "watch" will iteratively run every couple of seconds. Thanks.

[@ramalingamvarkala1571]

This video is very heplful.Can u make a video on blue/green deployment

[@justmeandopensource]

Hi Ramalingam, thanks for watching this video and suggesting a topic. I will add it to my list and get it recorded. Cheers.

[@VijayYadav-rc2is]

Sir i have a deployment with 3 replicas and i want to update the image with latest version.but i dont want to change the image name. Is it possible to do it with same image name?

[@tirthadeeproy1200]

could you please make a tutorial on DB schema migration (like adding new table and change column to existing table) in kubernetes rolling update

[@naveenkumar-jy3tw]

thanks.

[@justmeandopensource]

You are welcome Naveen!

[@pavanrajgupta641]

Hi , Thanks for posting videos, I have one question. What is the difference between $ kubectl patch deployment and Rolling update Strategy ?

[@justmeandopensource]

HI Pavan, thanks for watching. These two are different concept. I am not sure why you want to relate these two. Kubectl patch command is used to update your deployment. For example you can use kubectl patch deployment to change the image of the pods, or any other settings. Rolling update strategy is a mechanism where you define like maxunavailable, maxsurge like settings when it comes to updating the deployments.

[@biahoi87]

Hi Venkat, how the kubenetes behave in the case that I push a new container image for a hot fix in between the pause and resume deployment ?

[@justmeandopensource]

Hi Quan, I haven't tested that scenario. Just go ahead and try that and let me know what happens. This is how you learn. Thanks.

[@andrewnhien9714]

It can be, you can edit your yaml file and apply it, then you need to run command "resume" like Venkat show in video. It will create new replicaSet like you define in yaml file.

[@ridan1111]

I love your videos. amazingjob. do you have anything like this for openshfit aswell?

[@justmeandopensource]

Hi Ridan, thanks for watching. Openshift is in my list. Haven't started yet. I might start it soon. Cheers.

[@carlosvasquez4567]

gold stuff

[@justmeandopensource]

Thanks for watching.

[@TheLittlemaster2009]

hi #just me and opensource,hope you are doing good !! do you have any video on liveness and readiness probes and if it is i can't find it in :(

[@justmeandopensource]

Hi Shoaib, thanks for watching this video. I didn't do a video on those trivial topics that I should have done. I will add it to my list. Can't guanrantee when it will be released as I have recorded videos for the next two months and a lot in my list of stuff to cover. Cheers.

[@Praveen-xx7iy]

Is it possible to do a rolling update with a single replica without downtime?

[@justmeandopensource]

Hi Praveen, thanks for watching this video. Even when you have just one replica, when you update it, it should create a new pod and only when its ready the replication controller will terminate the old pod. The replication controllers job is to maintain 1 pod at any time. So you should have a zero downtime with that. Cheers.

[@anonymous9048]

Great video!! Just wanted to ask, by specifying maxSurge=2 and maxUnavailable=2, would the update mechanism be similar to a Blue Green deployment? Thanks

[@justmeandopensource]

Hi, Thanks for watching this video and taking time to comment. Blue green deployment requires you to maintain two identical environments, where you roll out new versions to identical environment and once verified traffic will switched to the new environment. And if something goes wrong, traffic can be switched back to the old environment. Thats a general idea behind blue green deployment. Whereas here, we are not doing any such logic. Instead we are rolling out new versions in the same environment gradually. Maxsurge and max unavailable can be chosen as per requirement and capacity. I just chose those numbers out of the sky. Thanks.

[@anonymous9048]

@@justmeandopensource Oh. I see... But both update mechanisms cause 0 downtime, correct? Thanks for the quick response :)

[@justmeandopensource]

Yes. You are right. Cheers.

[@anonymous9048]

@@justmeandopensource Perfect. Thanks!

[@justmeandopensource]

@@anonymous9048 You are welcome. Cheers

[@gautamsoni17990]

Hey Venkat, I have one doubt, when i deploy my deployment with rollingupdate with maxsurge: 0, maxunavailable: 1, minreadyseconds: 5 and revisionhistorylimit: 10 When i run the deploy, my 4 container created and start running fine but when i change the image tag and then apply the same yaml file, only 1 container got updated and i have waited for 5 minutes but nothing happened on 2nd container??? I have used first nginx:alpine and then used nginx:1.15.1

[@gautamsoni17990]

Please let me know some use case of maxsurge and maxunavailable as these are not clear yet

[@gautamsoni17990]

i have tried to do multiple things with this yaml file, and what i have found is if i remove the minReadySeconds option than my rollout will work perfectly and all container start terminating and all new container started together... but i want the same like your video, i want to update one by one... Could you please explain little more??

[@gautamsoni17990]

When i am using the microk8s on ubuntu then this issue not coming.... fyi

[@gautamsoni17990]

I dont know what happened but now it is start working... i havent changed anything... i just rebooted the server and then try the same.... and i am happy that it is working now

[@justmeandopensource]

@@gautamsoni17990 Perfect. That save me some time. Thanks for the confirmation. Cheers.

[@akshay_metgud]

Thank you very much venkat.. You rocks. I have some queries with respect to the below error : Im gettting same error with minikube as well as in ubuntu VM. Could you please help on the below? NAME STATUS MESSAGE ERROR controller-manager Unhealthy Get "127.0.0.1:10252/healthz": dial tcp 127.0.0.1:10252: connect: connection refused scheduler Unhealthy Get "127.0.0.1:10251/healthz": dial tcp 127.0.0.1:10251: connect: connection refused etcd-0 Healthy {"health":"true"}

[@nah0221]

Hello awesome man .. could you make a tutorial for rolling out & back an application contains database. So, we can see how to manage database migration ?!

[@justmeandopensource]

Hi, thanks for watching. Database layer is kept separate independently of the front end so that we can update or scale the front end. Database migrations must be handled separately to front end.

[@realthought2262]

hello , at 25:08 you modified the changes cause of revision 2 but what if i want to change it for revision 1 , coz you haven't had mentioned any revision no so i get that it will change the last revision by default if i run : kubectl annotate deployment nginx-deploy kubernetes.io/change-cause="Updated to revision 1.14.2"

[@justmeandopensource]

hi, thanks for watching. I am not sure if thats possible. By default annotate updates the last revision with the supplied change cause. For any new rolling updates you can use --record flag or you can update the manifests directly. I will have to test this. kubernetes.io/docs/concepts/workloads/controllers/deployment/#checking-rollout-history-of-a-deployment

[@realthought2262]

@@justmeandopensource thx for the reply, I will also try if found something will let you know as well.

[@justmeandopensource]

@@realthought2262 that will be useful to know thanks.