Thanks Elton! This was fantastic quick walkthrough. I would suggest making more videos around different aspects of docker, and publishing them. Would be very useful to the community. I have seen your many other courses, and must say your explanation style is superb.
I found this video EXTREMELY helpful. Completely making swarm approachable. I'd love to see it updated with the current docker version. But thanks for doing this Elton.
This is a very clear and concise walk-through of an introduction to docker swarms. It is especially effective with the combination of diagrams and command line. Well done!
This is extraordinary! I am experimenting now to use Docker swarm to distribute 5G mobile networking SDN nodes, and simulate DoS attack on mobile infrastructure with prevention mechanisms for my master (and later PhD thesis). The redundancy thing (if a node fails, other prevails) is a godly thing! This can be the Chuck Norris of mobile network security. Great video :) Cheers!
nice video Elton ---- u have great talent and I would request you to make more videos on all the knowledge you have ---which is help full for others like me ______thanks elton
thanks. can you also show how to add --constraint to docker service? i've tried adding --label name=node1 to the daemon and --constraint node.labels.name==node1 or --constraint node==node1, with no luck.
Will the swarm load balancer route traffic like a proxy does? In a real life scenario, one seldom have only one service/site that needs scaling. I have several different services/sites that needs to be exposed to the external network. I'm used to working with IIS and nginx, where one would tell the webserver to listen to port 80/443 and then configure the webserver to send the traffic to specific sites based on the incoming URL. There is no restriction on how many sites that can be exposed behind those external ports. However, when I try to run several sites/services on my swarm, either all URLs are routed to the same site(container) or none of them work at all. When I try to start several services with the flag -p 80:8080(which would be external port 80 mapped to port 8080 on the container) and then start another service with -p 80:3000, I get a conflict. I have also tried the -e VIRTUAL_PORT flag, to let the swarm figure it out itself, but then the services doesn't respond at all. Will I have to configure an external load balancer to direct my traffic to, say like www.example1.com -> nginx proxy www.example2.com-> nginx proxy then from there shoot all my traffic to the swarm manager on the correct container port? How would that be affected when scaling up or down? I have found no good resources covering this topic. There is always just one site being scaled up or down, which indeed is quite simple, but noone seem to cover more complex scenarios. Any input on how this could be approached will be highly appreciated.
Dude this has been so helpful thank you!!! I have been messing up trying to create docker machine swarm nodes and trying to use swarm that way but the information that the engine and the docker swarm commands are different has had me running in circles!! Although i dont see a node column when im running the docker task ls command.
Very good tutorial, thank you. However, the load-balancing needs further explanation as it lacks a final example of connection from the browser when replicas are spun up. Can you get rid of the initial external balancer and simply connect to the first one node expecting the swarm to re-route the request to the nodes where the tasks are running?
Hi, I was trying this in Amazon EC2. I want to ask what are the ports that must be opened for routing mesh to work exactly. It does not work when I just opened the ports 2377, 7946, 4789, 22 and 80. It then magically worked after opening ALL ports to public.
Thank You Elton this is a nice simple and clear tutorial as introduction to docker swarm ! i have a question what if an node with container created already with docker run command ,can swarm recognize those containers ,can i replicate theme to the others node ? thanks in advance
In your video, you shutdown swarm-03 to simulate a failure. What happens when swarm-03 boots again and joins the cluster? Would the tasks be restored to swarm-03, or would they continue to run on the other swarm nodes?
Thx a lot for your sharing, Captain! I'm new in swarm and I'm using AWS EC2 as nodes for swarm. I created 3 nodes including swarm manager and I cannot join swarm like you did in the video. each node has its inner IP and a public IP different from it. as your nodes are all in the same subnet, what can I do to join swarm? should I need token to join and which IP should I use?
Thanks Elton! for the simple yet amazing explanation.I have a question regarding manager node, what if the manager node goes down is there any way it can delegate the responsibility to some other node.
I see the benefits of the swarm for business use. For home use, maybe not so much. Could I run a tvheadend docker in the swarm? It can only handle a certain number of streams on the pi before the cpu maxes out. It would be great if I could install it in a swarm and if one server has the maximum number of streams it would spawn another instance of tvheadend on another node
How do you use docker join without token? I try but without sucess. "Error response from daemon: rpc error: code = 3 desc = A valid join token is necessary to join this cluster".
In the new version of Docker, the tokens are must. So when you will run the "docker service init" you will get the join command in result, which you can hit in your other nodes.
Hello I'm when I'm trying to join a docker node to the docker swarm I'm getting an error: docker swarm join x.x.x.x:2377 Error response from daemon: rpc error: code = Unavailable desc = all SubConns are in TransientFailure, latest connection error: connection error: desc = "transport: Error while dialing dial tcp x.x.x.x:2377: connect: no route to host"
Thank you for this great presentation! I have two questions please: how can i connect different hosts machines into the same network ? and how can i say that this particular machine is connected to this overlay network (if overlay network is the solution) ?
If i get 10 physical servers running 10 replicas of website and 5 of them fail. What do i get from launching the remaining 5 replicas on the remaining 5 servers? It will only increase load on the last standing servers by running more instances of webserver. If i don't do anything, the remaining webserver will just handle the load anyway (since it's the same physical host) without more containers eating the ram.
Nice video. What happen if an unknown client in the subnet launches the join command? It will ba added automatically to the cluster? Is there a possibility to confirm the adding from the manager?
What happens if the manager goes down instead of a worker? will another worker automatically take up being the manager or will the the whole swarm go down?
Hi, I am running docker in swarm mode in EC2 instances. When I am running my webservice as a docker service in manager node, I am not able to view the same from client node(using client public IP in browser). Is there any way to fix this ?
May I suggest you to have multiple tabs open in the terminal, so we dont have to keep ssh-ing multiple times. It was a perfect introduction otherwise. Thank you.
hi, this looks amazing and easy to setup compared to DCOS or K8S. What happens if we want to start two types of web sites both on port 80. Say I host two web sites with different host names.
Thank you Elton! That was really cool! :-) Can you tell what happens if node-00 (the manager node) fails? Does docker swarm reassign another node as the new manager or would the whole swarm go down?
Hi Andy, glad you liked it. In this case there's only one manager, so if you lose it then you can't manage the swarm until it comes back. In production you'd have 3 or 5 managers, one is the active leader, and the others elect a new leader if it goes down.
Your pacing is perfect. I suggest you making more videos of these kind. Your pacing is just right.
I watched alot of tutorial and this is the only one I understand. Indeed a good teacher makes a huge difference. Thank you.
You're an awesome instructor. I could listen for hours. I don't even have to pause and rewind. Got it right away.
Thanks Elton! This was fantastic quick walkthrough. I would suggest making more videos around different aspects of docker, and publishing them. Would be very useful to the community. I have seen your many other courses, and must say your explanation style is superb.
Thanks for that Arjun, glad you found it useful. I am planning more, just need to find the time.
This is exactly I was looking to understand. Thankyou very very much.
I found this video EXTREMELY helpful. Completely making swarm approachable. I'd love to see it updated with the current docker version. But thanks for doing this Elton.
This is the best technique intro I've ever seen.
Very good explanation and fantastic instructor. I understood the swarm concept by just watching 12 minutes session. That's Great!
You explained it so simply. Thanks a lot.
This is a very clear and concise walk-through of an introduction to docker swarms. It is especially effective with the combination of diagrams and command line. Well done!
I loved every bit of it. Great video !
Excellent and concise! I especially appreciate the summary at the end. Please more Docker videos for us newbs.
That's a nice and simply demo and explanation
EXCELLENT...!! Explanation was flawless. Easy to follow
Excellent, quick, and easy to understand demo! Thank you!
Great video about Swarms. Understood it in one go!!!
Excellent demonstration. More of these please!
This is extraordinary! I am experimenting now to use Docker swarm to distribute 5G mobile networking SDN nodes, and simulate DoS attack on mobile infrastructure with prevention mechanisms for my master (and later PhD thesis). The redundancy thing (if a node fails, other prevails) is a godly thing! This can be the Chuck Norris of mobile network security. Great video :) Cheers!
Great presentation, short and concise.
Thanks Elton.
Simply perfect explanation, I was looking around how to learn and understand swarm and you teach it in a simple and consise way
This is a very clean and simple explanation and demo for understanding this! Thank you so much!
You're an awesome instructor. Simple but in detailed. Thank you so muchhh!!!
Amazing Tutorial! Understood every part of it! Well done Elton!
Your video has provided a clear structure and awesome. Thank you so much!
Please make some more videos on docker
i loved how you explained
Excellent Video. Thanks for your effort Elton.
can't comment more than this.. just a perfect one.. subscribed and liked..
Awesome Elton, What great walk through. I Appreciate this very much. Thanks!
very nice and to the point tutorial. grate job man!
nice video Elton ---- u have great talent and I would request you to make more videos on all the knowledge you have ---which is help full for others like me ______thanks elton
This was an amazing video and so clear.
A nice quick walkthrough, Thanks!
Everything is so perfect in this video ..fantastic job ..thanks a ton :)
Thank you for posting this. Perfection.
Thank you !! Really clear and useful
Thanks a lot for this video! Very well explained.
thanks Elton! this video helps me alot
Very well presented... Perfect!
Thanks Elton! Well explained video!
Nice job with walk-through. Keep it up.
You teach very well, thank you
Awesome explanation!
awesome narration
Very nice....simple n stright
thanks. can you also show how to add --constraint to docker service? i've tried adding --label name=node1 to the daemon and --constraint node.labels.name==node1 or --constraint node==node1, with no luck.
Will the swarm load balancer route traffic like a proxy does? In a real life scenario, one seldom have only one service/site that needs scaling. I have several different services/sites that needs to be exposed to the external network. I'm used to working with IIS and nginx, where one would tell the webserver to listen to port 80/443 and then configure the webserver to send the traffic to specific sites based on the incoming URL. There is no restriction on how many sites that can be exposed behind those external ports.
However, when I try to run several sites/services on my swarm, either all URLs are routed to the same site(container) or none of them work at all. When I try to start several services with the flag -p 80:8080(which would be external port 80 mapped to port 8080 on the container) and then start another service with -p 80:3000, I get a conflict. I have also tried the -e VIRTUAL_PORT flag, to let the swarm figure it out itself, but then the services doesn't respond at all.
Will I have to configure an external load balancer to direct my traffic to, say like
www.example1.com -> nginx proxy
www.example2.com-> nginx proxy
then from there shoot all my traffic to the swarm manager on the correct container port? How would that be affected when scaling up or down? I have found no good resources covering this topic. There is always just one site being scaled up or down, which indeed is quite simple, but noone seem to cover more complex scenarios. Any input on how this could be approached will be highly appreciated.
Excellent content
Dude this has been so helpful thank you!!! I have been messing up trying to create docker machine swarm nodes and trying to use swarm that way but the information that the engine and the docker swarm commands are different has had me running in circles!! Although i dont see a node column when im running the docker task ls command.
Thank you for this Elton ! Crystal clear. Will I have to get more infos regarding load balancing now .. thanks
Great video! Thanks
Good pace. .. will be great to know how the swarm compares with Kubernetes..
very useful, great presentation
That was great. You should post more videos
It is a very good video.
What if the manager goes down? will the worker node is promoted to be a manager automatically ?
Rick Wong I’m limited in my knowledge but it seems common practice to add a few managers based on the size of your swarm
Hi! How can I load balance WebSockets? They need sticky sessions, so docker swarm is unable to handle simple WebSockets?
🤔
Very good tutorial, thank you. However, the load-balancing needs further explanation as it lacks a final example of connection from the browser when replicas are spun up. Can you get rid of the initial external balancer and simply connect to the first one node expecting the swarm to re-route the request to the nodes where the tasks are running?
Hi, I was trying this in Amazon EC2. I want to ask what are the ports that must be opened for routing mesh to work exactly. It does not work when I just opened the ports 2377, 7946, 4789, 22 and 80.
It then magically worked after opening ALL ports to public.
Found my answer!
TCP 2377
TCP 4789
TCP 7946 / UDP 7946
Thank You Elton this is a nice simple and clear tutorial as introduction to docker swarm !
i have a question what if an node with container created already with docker run command ,can swarm recognize those containers ,can i replicate theme to the others node ? thanks in advance
Amazing explanation!!
In your video, you shutdown swarm-03 to simulate a failure. What happens when swarm-03 boots again and joins the cluster? Would the tasks be restored to swarm-03, or would they continue to run on the other swarm nodes?
Thanks a lot!!! Great stuff!
Thx a lot for your sharing, Captain! I'm new in swarm and I'm using AWS EC2 as nodes for swarm. I created 3 nodes including swarm manager and I cannot join swarm like you did in the video. each node has its inner IP and a public IP different from it. as your nodes are all in the same subnet, what can I do to join swarm? should I need token to join and which IP should I use?
Nice talk Elton. Have you uploaded the slides somewhere?
Thanks Elton! for the simple yet amazing explanation.I have a question regarding manager node, what if the manager node goes down is there any way it can delegate the responsibility to some other node.
can you run two different services, that both use the same port, on one swarm?
If you have a Proxy Routing to the services sure. Take a look at jwilder/nginx-proxy
Thanks a lot @Elton Stoneman! One question tough: What happens id the load balancer fails? Can docker/docker swarm also offer a solution for this?
I see the benefits of the swarm for business use. For home use, maybe not so much. Could I run a tvheadend docker in the swarm? It can only handle a certain number of streams on the pi before the cpu maxes out. It would be great if I could install it in a swarm and if one server has the maximum number of streams it would spawn another instance of tvheadend on another node
is the Inner Load Balancing kinda soft one like zookeeper does? or embedded a nginx?
Awesome video. But what if the manager node goes down? Do other workers node elect a new manager?
Awesome video
Thanks Elton,
I wonder what happens if manager itself goes down for some awry unknown reason 🙈
I believe Managers can also belong to a group called Raft. Not entirely sure as I'm just learning this myself
How do you use docker join without token? I try but without sucess. "Error response from daemon: rpc error: code = 3 desc = A valid join token is necessary to join this cluster".
In the new version of Docker, the tokens are must. So when you will run the "docker service init" you will get the join command in result, which you can hit in your other nodes.
Hello I'm when I'm trying to join a docker node to the docker swarm I'm getting an error:
docker swarm join x.x.x.x:2377
Error response from daemon: rpc error: code = Unavailable desc = all SubConns are in TransientFailure, latest connection error: connection error: desc = "transport: Error while dialing dial tcp x.x.x.x:2377: connect: no route to host"
Thank you for this great presentation!
I have two questions please: how can i connect different hosts machines into the same network ? and how can i say that this particular machine is connected to this overlay network (if overlay network is the solution) ?
is the loadbalancer you set up is just anathor machine, which doesn't have the service running and can reroute the requests?
Thank you so much, very useful :D
Really a good video
If i get 10 physical servers running 10 replicas of website and 5 of them fail. What do i get from launching the remaining 5 replicas on the remaining 5 servers? It will only increase load on the last standing servers by running more instances of webserver. If i don't do anything, the remaining webserver will just handle the load anyway (since it's the same physical host) without more containers eating the ram.
Nice video.
What happen if an unknown client in the subnet launches the join command? It will ba added automatically to the cluster? Is there a possibility to confirm the adding from the manager?
Great video!! Would be possible to set an autoscale config for docker swam?
Thanks, if I want to cluster the server in different servers that maybe in different networks what should I do?
What happens if the manager goes down instead of a worker? will another worker automatically take up being the manager or will the the whole swarm go down?
hi elton how can i schedule service on a swarm from a non manager machine like a jenkins machine
Hi, great video. May I ask which tool did you use to build the infographics?
Thanks.
Bernardo Corrêa PowerPoint!
Nice video thanks!!!
Really helpful
Hi, I am running docker in swarm mode in EC2 instances. When I am running my webservice as a docker service in manager node, I am not able to view the same from client node(using client public IP in browser). Is there any way to fix this ?
May I suggest you to have multiple tabs open in the terminal, so we dont have to keep ssh-ing multiple times. It was a perfect introduction otherwise. Thank you.
Thanks man, really helpful
Hey very new to docker swarm . If I install a container on manger and want the workers and manager to share cpu load is that possible?
hi,
this looks amazing and easy to setup compared to DCOS or K8S.
What happens if we want to start two types of web sites both on port 80. Say I host two web sites with different host names.
Hi Elton, Know your using nginx in the website image, but can show us the dockerfile to see how your configuring it. - thanks.
Thank you Elton! That was really cool! :-)
Can you tell what happens if node-00 (the manager node) fails? Does docker swarm reassign another node as the new manager or would the whole swarm go down?
Hi Andy, glad you liked it. In this case there's only one manager, so if you lose it then you can't manage the swarm until it comes back. In production you'd have 3 or 5 managers, one is the active leader, and the others elect a new leader if it goes down.
Why would you create regular worker nodes instead of making every node a manager?
Where can we find your training .
Excellent. Thanks.
Fantastic.
Congrats! Thats a really cool and clear video!
So, at 11:00 : you shutdown one of the worker nodes. What will happen if the manager node is down?
I was a bit lost when you said docker swarm has it own load balancer and that you are also using azure load balancer???
DOCKS ROCKS
O Captain! My Captain! Nice vid! :)