TCE is great i think and i prefer it alot rather then the alternatives especially for on prem. The developers and managers are also very responsive and listen to feedback and it is rapidly evolving based on feedback in a much better way i believe then many other tools in the ecosystem
IMHO Rancher seems to be one level up. The possibility to manage self-hosted and managed clusters is great. Rancher API abstraction (crossplane kinda), Rancher fleet, etc put Rancher many steps ahead. Rancherctl is limited in terms of cli mgmt, but it's not the end of the world (maybe they have improved it on RKE2).
@Lucas Albuquerque i actually used to think the same till i got into really using tanzu. I still sometimes use rancher as the ui above TCE in some cases, as thats the beauty of conformant k8s. The api of rancher is i think terrible and makes no sense how it was documented. They also haved "moved" to using CAPI for management but are completely not aligned with the community. They built there own providers that do rke based provisioning, removed parts of capi, dont rely on the upstream infra providers if capi etc. Rancher seems to be doing to CAPI what openshift did to k8s. On one hand thats fine if it gives you a good experience but its not upstream aligned, it will get updates typically much later, and is much more prone to lock in. Im not a big fan of fleet actually and prefer flux which TCE includes as a package as well as kapp controller which also can do gitops which is pre installed in every cluster. Regarding managing self hosted and managed k8s clusters, this can already be done in a different tanzu tool called TMC. The free version of that SaaS offering has been announced already and is in beta right now. Also full management of managed clusters is being worked on in tanzu as well.
So a few notes: 1. The management cluster is managed via CAPI. The local docker image is a kind cluster that installs capi and in the demo case CAPA, then creates the mgmt cluster and then uses the clusterctl move functionality to move the mgmt capi resources to itself making it a self managed cluster. 2. The tanzu-package-repo-global namespace is what in carvel packaging is refered to as the global namespace. You can install a repo in any namespace making its packages namespace scoped. If it is installed in that namespace the packages are globally accesible in the cluster 3. The reason for grafana being old is that its the last version thats not agpl licensed. Other packages are updated frequently 4. The requirment for dependant packages is being worked on 5. There is a web ui coming very soon to the community. Tanzu mission control allow management of clusters and a free version is coming. Also a TCE specific UI is coming as well 6. Managed clusters is also being worked on currently. 7. Google support is lacking as the CAPI provider is only recently being revived but is planned to be added when the project itself is healthy 8. Update is currently via kubectl on the capi resources itself and is indeed a pain point 9. Autoscaler is only deployed at cluster creation time currently. 10. Auto installing packages is being worked on and is added in the next version of TCE. 11. Take a look at tanzu unmanaged clusters as well as they are a huge feature of TCE. 12. You can use any local k8s cluster as the bootstrap cluster using the --existing-bootstrap-cluster flag 13. The bitnami repo is being made available as packages soon as well, i already converted the entire repo to a package repo which gives you over 100 additional packages. 14. Would love to discuss some additional things i find awesome in tanzu with you if you are interested that you didnt go into in this video that make it awesome and unique 15. In terms of what you get beyond capi there is alot, authentication system, single click upgrade and much more.
This is just a superb presentation. Perfect amount of humour to lighting it up and at the same time a very clearly knowledgeable explication of complex items. The console projection is also amazing. Prefect readable and you still can your expressions with so important in a teaching environment.
Since you asked MULTUS CNI: Multus CNI is a container network interface (CNI) plugin for Kubernetes that enables attaching multiple network interfaces to pods. Yes some pods want multiple interfaces to offer segmented networking across different services they offer/use to/from network fabric.
thanks for sharing.... quick question.. which version are you using in the video? I noticed that your installer has an option to install using docker in the local machine .... But I tried with several versions i couldn't find that option....
For a few months now, I do not edit videos anymore. The "terminal effect" is done by a guy that does the editing/animations/sounds/etc. I do separate recordings of my face and screen so I'm guessing that he (the editor) is removing the terminal background and overlaying it on top of my face.
Great video with explanation on that platform, for which only vmware docs can be read but its kinda complex... if you can do more videos about kubernetes on tanzu it would be great
I am not sure how relavent tanzu in kubernetes times. Cloud foundry, cf push were ahead of k8s times. Brought virtualyzed deployment and orchestration via Bosh. Buy now k8s is the defacto container deployment, cloud foundry has no option but ditch Bosh, and embrace k8s via cf-for-k8s initiaitve. What remains constant from the old times is cf push developer experience. Dos it matter in current times?
CF was revolutionary but faced the same fate as many other pioneers faced. First iterations of ideas are often replaced by new ones based on experience of those. Something similar happened to Heroku, mesos, etc. Sometimes, it's not good to be the first or to come up with something too early. Today, we (the industry) settled on Kubernetes and now it's all about building the layers on top of it and VMWare might have a hard time letting go the past (CF).
I am fresh with kubernetes but heavy with experience with VMware. I tried open shift and tanzu and feel much better with red hat product. Company though is vendor locked with VMware so I do not have much choice:(
Also the local mgmt cluster is spun up using KIND by default and it runs cluster api on the kind cluster. IIRC you can point it to any cluster. WRT to managed Kubernetes cluster, try TMC it has a nice ui for most of this.
That would be great. Maybe we can do some material together. How about we start with a chat? If that sounds good, please pick any time that works for you from calendly.com/vfarcic/meet
I haven't tried (yet) Spectro Cloud. Adding it to my TODO list... I wasn't very impressed with Kubespray in the past, but that was a while ago when it was much greener than it is today. So, I have to see it again before saying anything about it.
I haven't (yet) spent much time with Charmed so I cannot (yet) compare it with Tanzu. Microk8s, on the other hand, is a different category. It could be, for example, compared with k3s.
Promoting their own open source projects (which underlies their commercial products) and providing an opinionated path through the big cloud native ecosystem? And they have a lot invested in things like carvel, cartographer, kpack etc all related to building, CI/CD etc so to providing some app deployment functionality makes a lot of sense. Giving their Tanzu CE users a glimpse of something that they might want to get commercial support for in the future
Actually It is not exactly like on-prem. ClusterAPI is a cluster control plane, it will automatically ensure instances are healthy, help with rolling upgrades and more. It is close to providing a QoS close to EKS if you can ensure your management cluster is highly-available. The only thing you do not get with ClusterAPI is guaranteed SLA, this part is on you :)
ClusterAPI is great. Still, if one is running in AWS, there should at least be an option to manage it as EKS. I'm not saying that everyone using AWS should be using EKS, but many do and EKS support (among others) feels like should be part of Tanzu (just as it's part of Rancher).
Adding it to my TODO list... :) I see that you're working for Cognit so it might make sense to have a quick chat about Castor. If that makes sense, please send me a DM on Twitter (@vfarcic) or LinkedIn (www.linkedin.com/in/viktorfarcic/).
What do you think about Tanzu? Is it a better or worse choice than other similar platforms (e.g., Rancher, OpenShift, etc.)?
It’s expensive
The community edition is free buy the question is whether it is good enough of stripped down to the level that the enterprise edition is a must
TCE is great i think and i prefer it alot rather then the alternatives especially for on prem. The developers and managers are also very responsive and listen to feedback and it is rapidly evolving based on feedback in a much better way i believe then many other tools in the ecosystem
IMHO Rancher seems to be one level up. The possibility to manage self-hosted and managed clusters is great. Rancher API abstraction (crossplane kinda), Rancher fleet, etc put Rancher many steps ahead. Rancherctl is limited in terms of cli mgmt, but it's not the end of the world (maybe they have improved it on RKE2).
@Lucas Albuquerque i actually used to think the same till i got into really using tanzu.
I still sometimes use rancher as the ui above TCE in some cases, as thats the beauty of conformant k8s. The api of rancher is i think terrible and makes no sense how it was documented. They also haved "moved" to using CAPI for management but are completely not aligned with the community. They built there own providers that do rke based provisioning, removed parts of capi, dont rely on the upstream infra providers if capi etc. Rancher seems to be doing to CAPI what openshift did to k8s. On one hand thats fine if it gives you a good experience but its not upstream aligned, it will get updates typically much later, and is much more prone to lock in.
Im not a big fan of fleet actually and prefer flux which TCE includes as a package as well as kapp controller which also can do gitops which is pre installed in every cluster.
Regarding managing self hosted and managed k8s clusters, this can already be done in a different tanzu tool called TMC. The free version of that SaaS offering has been announced already and is in beta right now. Also full management of managed clusters is being worked on in tanzu as well.
So a few notes:
1. The management cluster is managed via CAPI. The local docker image is a kind cluster that installs capi and in the demo case CAPA, then creates the mgmt cluster and then uses the clusterctl move functionality to move the mgmt capi resources to itself making it a self managed cluster.
2. The tanzu-package-repo-global namespace is what in carvel packaging is refered to as the global namespace. You can install a repo in any namespace making its packages namespace scoped. If it is installed in that namespace the packages are globally accesible in the cluster
3. The reason for grafana being old is that its the last version thats not agpl licensed. Other packages are updated frequently
4. The requirment for dependant packages is being worked on
5. There is a web ui coming very soon to the community. Tanzu mission control allow management of clusters and a free version is coming. Also a TCE specific UI is coming as well
6. Managed clusters is also being worked on currently.
7. Google support is lacking as the CAPI provider is only recently being revived but is planned to be added when the project itself is healthy
8. Update is currently via kubectl on the capi resources itself and is indeed a pain point
9. Autoscaler is only deployed at cluster creation time currently.
10. Auto installing packages is being worked on and is added in the next version of TCE.
11. Take a look at tanzu unmanaged clusters as well as they are a huge feature of TCE.
12. You can use any local k8s cluster as the bootstrap cluster using the --existing-bootstrap-cluster flag
13. The bitnami repo is being made available as packages soon as well, i already converted the entire repo to a package repo which gives you over 100 additional packages.
14. Would love to discuss some additional things i find awesome in tanzu with you if you are interested that you didnt go into in this video that make it awesome and unique
15. In terms of what you get beyond capi there is alot, authentication system, single click upgrade and much more.
The mgmt capi resources are in the tkg-system namespace by the way
Those are amazing notes. Thanks a ton.
... and yes. I'd love to chat with you about tanzu. Can you dm me on twitter or linkedin?
Your channel is a brilliant resource and the best youtube videos on Tanzu technology!
This is just a superb presentation.
Perfect amount of humour to lighting it up and at the same time a very clearly knowledgeable explication of complex items.
The console projection is also amazing. Prefect readable and you still can your expressions with so important in a teaching environment.
Since you asked MULTUS CNI: Multus CNI is a container network interface (CNI) plugin for Kubernetes that enables attaching multiple network interfaces to pods. Yes some pods want multiple interfaces to offer segmented networking across different services they offer/use to/from network fabric.
THIS VIDEO IS AMAZING. THANK YOU SO MUCH!!!
thanks for sharing.... quick question.. which version are you using in the video? I noticed that your installer has an option to install using docker in the local machine .... But I tried with several versions i couldn't find that option....
I do not remember the exact version. It should have been the latest at the time i created that video so it's quite old now.
I like this. I have to check out your other material. Do you collaborate with anyone on this?
I do those types of videos alone. There are live AMA sessions on this channel that often include guests.
Hi victor, good video as always, btw how to make video like you? place terminal front of your video?
For a few months now, I do not edit videos anymore. The "terminal effect" is done by a guy that does the editing/animations/sounds/etc. I do separate recordings of my face and screen so I'm guessing that he (the editor) is removing the terminal background and overlaying it on top of my face.
Great video with explanation on that platform, for which only vmware docs can be read but its kinda complex... if you can do more videos about kubernetes on tanzu it would be great
I am not sure how relavent tanzu in kubernetes times. Cloud foundry, cf push were ahead of k8s times. Brought virtualyzed deployment and orchestration via Bosh. Buy now k8s is the defacto container deployment, cloud foundry has no option but ditch Bosh, and embrace k8s via cf-for-k8s initiaitve. What remains constant from the old times is cf push developer experience. Dos it matter in current times?
CF was revolutionary but faced the same fate as many other pioneers faced. First iterations of ideas are often replaced by new ones based on experience of those. Something similar happened to Heroku, mesos, etc. Sometimes, it's not good to be the first or to come up with something too early.
Today, we (the industry) settled on Kubernetes and now it's all about building the layers on top of it and VMWare might have a hard time letting go the past (CF).
TAP from vmware which in the oss world of tce is called app toolkit gives a k8s native cf push like experience which is great
I am fresh with kubernetes but heavy with experience with VMware. I tried open shift and tanzu and feel much better with red hat product. Company though is vendor locked with VMware so I do not have much choice:(
Also the local mgmt cluster is spun up using KIND by default and it runs cluster api on the kind cluster. IIRC you can point it to any cluster.
WRT to managed Kubernetes cluster, try TMC it has a nice ui for most of this.
In that video, I limited the scope to OSS so I did not include TMC. I have it on my TODO list for a separate video.
Victor, how about Anthos and EKS-A?
I have both on my TODO list. I can't promise when they'll come, but I can say that they are coming.
@@DevOpsToolkit Victor, if you want to chat about EKS-A, I work on the team and can help answer any questions. Love your videos, keep it up!
That would be great. Maybe we can do some material together.
How about we start with a chat? If that sounds good, please pick any time that works for you from calendly.com/vfarcic/meet
How do you think that compares to Spectro Cloud and/or Kubespray?
I haven't tried (yet) Spectro Cloud. Adding it to my TODO list...
I wasn't very impressed with Kubespray in the past, but that was a while ago when it was much greener than it is today. So, I have to see it again before saying anything about it.
Hi, thanks a lot for great video. It would be nice to see something similar related to openshift. Please do you have it on the list? Thanks.
I already have it on my TODO list. The goal was to cover the "big three" (Rancher, Tanzu, and OpenShoft). The first two are done so OpenShift is next.
What about Charmed and MicroK8s?
I haven't (yet) spent much time with Charmed so I cannot (yet) compare it with Tanzu. Microk8s, on the other hand, is a different category. It could be, for example, compared with k3s.
Why all this additional package foo per commercial k8s distribution
Promoting their own open source projects (which underlies their commercial products) and providing an opinionated path through the big cloud native ecosystem? And they have a lot invested in things like carvel, cartographer, kpack etc all related to building, CI/CD etc so to providing some app deployment functionality makes a lot of sense. Giving their Tanzu CE users a glimpse of something that they might want to get commercial support for in the future
Actually It is not exactly like on-prem. ClusterAPI is a cluster control plane, it will automatically ensure instances are healthy, help with rolling upgrades and more. It is close to providing a QoS close to EKS if you can ensure your management cluster is highly-available.
The only thing you do not get with ClusterAPI is guaranteed SLA, this part is on you :)
ClusterAPI is great. Still, if one is running in AWS, there should at least be an option to manage it as EKS. I'm not saying that everyone using AWS should be using EKS, but many do and EKS support (among others) feels like should be part of Tanzu (just as it's part of Rancher).
hi, please give a shot to "castor k8s"
Adding it to my TODO list... :)
I see that you're working for Cognit so it might make sense to have a quick chat about Castor. If that makes sense, please send me a DM on Twitter (@vfarcic) or LinkedIn (www.linkedin.com/in/viktorfarcic/).
why use this mess in place of standard k8s?
Tanzu is a standard k8s. It's not about which k8s to use but how to manage k8s clusters (infra, apps, upgrades, etc.).
Farcic please create full series on openshift
I'll do my best to do it.
I would really like to see rancher vs openshift. You could add Tanzu if you like :-)
First I'll do a video about Open Shift and, after that, a comparison between the three.
Came expecting Victor to install WordPress. Disappointed.
RIP TCE.
waiting for Carvel!!