Lease API | Kubernetes
6:48
Рік тому
КОМЕНТАРІ
@jchink2708
@jchink2708 16 днів тому
😔wonderful session, so admire your calm voice and clear presentation
@jaimecarranza2509
@jaimecarranza2509 27 днів тому
This was a nice one, thank you.
@arpitpattewar2513
@arpitpattewar2513 27 днів тому
I am not able to read yaml files
@siddharthshrivastava229
@siddharthshrivastava229 Місяць тому
At least make the command visible in your video
@yogithakakarla1716
@yogithakakarla1716 Місяць тому
Give us link how to download Prometheus using helm Give us those commands.. or office document
@jmmtechnology4539
@jmmtechnology4539 Місяць тому
Hey! The commands can be found here: artifacthub.io/packages/helm/prometheus-community/kube-prometheus-stack
@jasonnwakaeze1683
@jasonnwakaeze1683 Місяць тому
How do i configure the nodes to respect these conditions set, because it just exceeds 100% then times out. I have to come drain it manually. I use EKS
@isi24.8
@isi24.8 2 місяці тому
Thank you !! Really well explained
@carmiwisemon740
@carmiwisemon740 2 місяці тому
Thank you. This was a clear explanation of the subject. Who creates the StorageClass the Developer or the Admin?
@yashpawse
@yashpawse 2 місяці тому
Thanks for keeping it short and up to the point. Got it very well :)
@user-tt8td7iv3z
@user-tt8td7iv3z 2 місяці тому
i drained my master node now after uncordon it doesnt work up remain in notready state. how am i to deal it
@OrionPronk
@OrionPronk 3 місяці тому
Thanks for this great video. If I could make a suggestion, please increase your terminal font size for recordings. It's a bit hard to read with a Macbook and hard to distinguish a = from a : even when squinting my eyes :D
@haskell3702
@haskell3702 3 місяці тому
2022 mac x86?
@booyaadotorg
@booyaadotorg 3 місяці тому
Thanks for this helpful video you've provided some great pointers as well useful tips (bash help aliases and functions).
@knopeace
@knopeace 3 місяці тому
This was great. Thanks! If mul
@r00ty
@r00ty 3 місяці тому
This is super confusing and misleading. You can't create a PVC without a Storage Class. When a PVC does not specify a storageClassName, the default StorageClass is used
@thisisnotanofficialemail-7872
@thisisnotanofficialemail-7872 3 місяці тому
After watching all you tuberss finally found someone who's explaining exactly like i was in want
@vladx3539
@vladx3539 3 місяці тому
Could you make a tad louder :)
@danishsiddiqui7321
@danishsiddiqui7321 4 місяці тому
VERY HELPFUL VEDIO !
@Shainmeng
@Shainmeng 4 місяці тому
Thank you for this video, man. I just got my CKA badge today.
@jmmtechnology4539
@jmmtechnology4539 4 місяці тому
That's awesome man congrats!
@anveshreddyoffofficial6581
@anveshreddyoffofficial6581 4 місяці тому
What's the reason for a lease to be used by node? For kube API to know the status of the node, can't kubelet of that node just send the requests directly to the API server instead of updating the lease? I can't see how the lease is helpful for nodes. Although I agree with you for distributed case we need leases.
@jmmtechnology4539
@jmmtechnology4539 4 місяці тому
Good question, my understanding is that the main reason for using the lease API is for performance, in particular for larger clusters. You can find a detailed explanation on why leases are now used for node heartbeats here: github.com/kubernetes/enhancements/tree/master/keps/sig-node/589-efficient-node-heartbeats
@anveshreddyoffofficial6581
@anveshreddyoffofficial6581 4 місяці тому
@@jmmtechnology4539 Thanks for the reply :). Will check that out!!
@anveshreddyoffofficial6581
@anveshreddyoffofficial6581 4 місяці тому
@@jmmtechnology4539 I did check it out and as you said it's because of the performance. They used to do this earlier but because of etcd going out of storage easily due to these update frequency being just 10s and it has to store this info. They moved away from it by introducing a lease for this. Great read!. Thanks a lot for the resource. Would love if you make more videos on these concepts explaining things internally, why they did what they did. Thanks a lot.
@ogeelueze2513
@ogeelueze2513 4 місяці тому
Excellent explanation . Thanks for putting this out
@jmlineb
@jmlineb 4 місяці тому
I don't think the controller manager can provision a new node by itself if one dies. I think it has to work with a cloud Kubernetes managed service to do so, like EKS or AKS or GKE.
@jmmtechnology4539
@jmmtechnology4539 4 місяці тому
Hey John, yes you're bang on with that. My apologies for the mistake here. I cover the cloud controller manager in a bit more detail in this video ukposts.info/have/v-deo/nIeVfoObrKiIpas.html and the information should be accurate.
@jmlineb
@jmlineb 4 місяці тому
Hello! Great series of videos, thanks. One reaction to this one. I don't think that Kubernetes by itself can replace a node that has gone down. For that I think it relies on a managed cloud service in which it is embedded, like EKS or AKS or GKE.
@SM-ok3sz
@SM-ok3sz 4 місяці тому
Does your keyboard owe you money?
@shadyashraf1031
@shadyashraf1031 5 місяців тому
I am currently finishing a cloud devops nano degree, your simple explanation is 100% better than any of the nanodegree's mentors
@shadyashraf1031
@shadyashraf1031 5 місяців тому
One of the best content creators, please don't stop making explicit videos. You are life changing
@nitishrawat98
@nitishrawat98 5 місяців тому
Hey there will you be uploading content to become kubernetes engineer.
@AnaisUrlichs
@AnaisUrlichs 5 місяців тому
Great video, where can I follow you? LinkedIn, Twitter?
@AnaisUrlichs
@AnaisUrlichs 5 місяців тому
WHERE CAN I FIND YOUR ONLINE 👀 (besides here) I just came across your videos!
@jmmtechnology4539
@jmmtechnology4539 5 місяців тому
Hey Anais, thanks for checking out my channel. I’m available on LinkedIn www.linkedin.com/in/jamie-mac-manus
@alexnovik
@alexnovik 6 місяців тому
do not do etcdctl snapshot status!!!! it will corrupt your snapshot
@narigina6414
@narigina6414 6 місяців тому
21:50 didn't know this cheat. Thank you for sharing
@sanchayana2007
@sanchayana2007 6 місяців тому
It deserves more views
@user-iv8kp7rv7r
@user-iv8kp7rv7r 6 місяців тому
Is it possible for freshers to take this exam ..? Or prior experience is necessary.
@yunusemregormus201
@yunusemregormus201 7 місяців тому
thanks for fabulous content
@yunusemregormus201
@yunusemregormus201 7 місяців тому
thanks for fabilous content
@Roman-ec8ns
@Roman-ec8ns 7 місяців тому
Thanks for your efforts!
@hradayeshshukla1326
@hradayeshshukla1326 8 місяців тому
Good one, thanks for explaining it in detail.👍
@user-ji8sw5sw8y
@user-ji8sw5sw8y 8 місяців тому
Just awesome mate !! kudos
@LeightonSpicer
@LeightonSpicer 8 місяців тому
Thanks John - Great video. Just need to put some effort in now.
@shmuelsigler2918
@shmuelsigler2918 9 місяців тому
im gonna use my laptop for the exam, do you know if i can use 2 monitors? great guide, i realy found it helpful.
@jmmtechnology4539
@jmmtechnology4539 9 місяців тому
As far as I’m aware additional monitors are not allowed but you should double check that on the pre exam guide once you have purchased the exam.
@randayho
@randayho 9 місяців тому
great idea on the function to change to the namespace. i'm also going to put an alias called "cn" so i can see the current namespace i'm in, i think that would be helpful also
@hhue13
@hhue13 10 місяців тому
Can I send an alert to multiple receivers as well? For example I have a receiver called NYC and another receiver PROD. If I have an alert with the label city=nyc and env=prod I'd like to send this alert to the receiver NYC and to receiver PROD. How could I achieve that?
@sohaibmsalha6742
@sohaibmsalha6742 10 місяців тому
Thanks man, subscribed!
@jmmtechnology4539
@jmmtechnology4539 10 місяців тому
Thanks!
@mdaverde
@mdaverde 10 місяців тому
Great video, why do you think the scheduler lives on its own rather than inside the controller manager? After all, it seems to fit into the job of reconciling the state of pods by assigning a node. Seems like this logic could fit inside a Pod controller of some sort
@jmmtechnology4539
@jmmtechnology4539 4 місяці тому
Hi Milan, that's a great question, I believe the scheduler and controllers have sufficiently distinct characteristics and responsibilities such that those who designed and implemented kubernetes thought they should remain separate. For example, the controllers are generally responsible for maintaining cluster state whereas the scheduler has the specific job of placing pods on the correct node. This respects the patterns of decoupling and modularity that are ubiquitous in software design.
@ashermanangan
@ashermanangan 10 місяців тому
thanks a lot ~ greetings from Philippines
@abdulshaikh6807
@abdulshaikh6807 10 місяців тому
But what about the auto scaling of the nodes, in that scenario as well we will face that issue ?
@programmer9809
@programmer9809 10 місяців тому
13:00
@RaviKolla
@RaviKolla 10 місяців тому
One of the best tutorials I have ever watched, the way you explain the information is pretty clear and direct. Thank you
@ZinnyAbraham
@ZinnyAbraham 10 місяців тому
Good content🎉
@PlacidoYT
@PlacidoYT 10 місяців тому
Hey man, how's the vm? Is it laggy? I do utilize k8s doc most of the times and copy codes from there.
@jmmtechnology4539
@jmmtechnology4539 10 місяців тому
The VM works nicely once your internet connection is good. If you're below 15Mb/sec you might start to run into lag.