[openstack-dev] [magnum]swarm + compose = k8s?

Hongbin Lu hongbin.lu at huawei.com
Sat Feb 13 03:00:04 UTC 2016


Thanks for sharing your insights. I gave it more thoughts. Maybe the goal can be achieved without implementing a shared COE. We could move all the master nodes out of user tenants, containerize them, and consolidate them into a set of VMs/Physical servers.

I think we could separate the discussion into two:

1.       Should Magnum introduce a new bay type, in which master nodes are managed by Magnum (not users themselves)? Like what GCE [1] or ECS [2] does.

2.       How to consolidate the control services that originally runs on master nodes of each cluster?

Note that the proposal is for adding a new COE (not for changing the existing COEs). That means users will continue to provision existing self-managed COE (k8s/swarm/mesos) if they choose to.

[1] https://cloud.google.com/container-engine/
[2] http://docs.aws.amazon.com/AmazonECS/latest/developerguide/Welcome.html

Best regards,

From: Guz Egor [mailto:guz_egor at yahoo.com]
Sent: February-12-16 2:34 PM
To: OpenStack Development Mailing List (not for usage questions)
Cc: Hongbin Lu
Subject: Re: [openstack-dev] [magnum]swarm + compose = k8s?


I am not sure that it's good idea, it looks you propose Magnum enter to "schedulers war" (personally I tired from these debates Mesos vs Kub vs Swarm).
If your  concern is just utilization you can always run control plane at "agent/slave" nodes, there main reason why operators (at least in our case) keep them
separate because they need different attention (e.g. I almost don't care why/when "agent/slave" node died, but always double check that master node was
repaired or replaced).

One use case I see for shared COE (at least in our environment), when developers want run just docker container without installing anything locally
(e.g docker-machine). But in most cases it's just examples from internet or there own experiments ):

But we definitely should discuss it during midcycle next week.


From: Hongbin Lu <hongbin.lu at huawei.com<mailto:hongbin.lu at huawei.com>>
To: OpenStack Development Mailing List (not for usage questions) <openstack-dev at lists.openstack.org<mailto:openstack-dev at lists.openstack.org>>
Sent: Thursday, February 11, 2016 8:50 PM
Subject: Re: [openstack-dev] [magnum]swarm + compose = k8s?

Hi team,

Sorry for bringing up this old thread, but a recent debate on container resource [1] reminded me the use case Kris mentioned below. I am going to propose a preliminary idea to address the use case. Of course, we could continue the discussion in the team meeting or midcycle.

Idea: Introduce a docker-native COE, which consists of only minion/worker/slave nodes (no master nodes).
Goal: Eliminate duplicated IaaS resources (master node VMs, lbaas vips, floating ips, etc.)
Details: Traditional COE (k8s/swarm/mesos) consists of master nodes and worker nodes. In these COEs, control services (i.e. scheduler) run on master nodes, and containers run on worker nodes. If we can port the COE control services to Magnum control plate and share them with all tenants, we eliminate the need of master nodes thus improving resource utilization. In the new COE, users create/manage containers through Magnum API endpoints. Magnum is responsible to spin tenant VMs, schedule containers to the VMs, and manage the life-cycle of those containers. Unlike other COEs, containers created by this COE are considered as OpenStack-manage resources. That means they will be tracked in Magnum DB, and accessible by other OpenStack services (i.e. Horizon, Heat, etc.).

What do you feel about this proposal? Let’s discuss.

[1] https://etherpad.openstack.org/p/magnum-native-api

Best regards,

From: Kris G. Lindgren [mailto:klindgren at godaddy.com]
Sent: September-30-15 7:26 PM
To: openstack-dev at lists.openstack.org<mailto:openstack-dev at lists.openstack.org>
Subject: Re: [openstack-dev] [magnum]swarm + compose = k8s?

We are looking at deploying magnum as an answer for how do we do containers company wide at Godaddy.  I am going to agree with both you and josh.

I agree that managing one large system is going to be a pain and pas experience tells me this wont be practical/scale, however from experience I also know exactly the pain Josh is talking about.

We currently have ~4k projects in our internal openstack cloud, about 1/4 of the projects are currently doing some form of containers on their own, with more joining every day.  If all of these projects were to convert of to the current magnum configuration we would suddenly be attempting to support/configure ~1k magnum clusters.  Considering that everyone will want it HA, we are looking at a minimum of 2 kube nodes per cluster + lbaas vips + floating ips.  From a capacity standpoint this is an excessive amount of duplicated infrastructure to spinup in projects where people maybe running 10–20 containers per project.  From an operator support perspective this is a special level of hell that I do not want to get into.   Even if I am off by 75%,  250 still sucks.

From my point of view an ideal use case for companies like ours (yahoo/godaddy) would be able to support hierarchical projects in magnum.  That way we could create a project for each department, and then the subteams of those departments can have their own projects.  We create a a bay per department.  Sub-projects if they want to can support creation of their own bays (but support of the kube cluster would then fall to that team).  When a sub-project spins up a pod on a bay, minions get created inside that teams sub projects and the containers in that pod run on the capacity that was spun up  under that project, the minions for each pod would be a in a scaling group and as such grow/shrink as dictated by load.

The above would make it so where we support a minimal, yet imho reasonable, number of kube clusters, give people who can't/don’t want to fall inline with the provided resource a way to make their own and still offer a "good enough for a single company" level of multi-tenancy.



>If you share resources, you give up multi-tenancy.  No COE system has the

>concept of multi-tenancy (kubernetes has some basic implementation but it

>is totally insecure).  Not only does multi-tenancy have to “look like” it

>offers multiple tenants isolation, but it actually has to deliver the



>I understand that at first glance a company like Yahoo may not want

>separate bays for their various applications because of the perceived

>administrative overhead.  I would then challenge Yahoo to go deploy a COE

>like kubernetes (which has no multi-tenancy or a very basic implementation

>of such) and get it to work with hundreds of different competing

>applications.  I would speculate the administrative overhead of getting

>all that to work would be greater then the administrative overhead of

>simply doing a bay create for the various tenants.


>Placing tenancy inside a COE seems interesting, but no COE does that

>today.  Maybe in the future they will.  Magnum was designed to present an

>integration point between COEs and OpenStack today, not five years down

>the road.  Its not as if we took shortcuts to get to where we are.


>I will grant you that density is lower with the current design of Magnum

>vs a full on integration with OpenStack within the COE itself.  However,

>that model which is what I believe you proposed is a huge design change to

>each COE which would overly complicate the COE at the gain of increased

>density.  I personally don’t feel that pain is worth the gain.

Kris Lindgren
Senior Linux Systems Engineer

OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request at lists.openstack.org<mailto:OpenStack-dev-request at lists.openstack.org>?subject:unsubscribe

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20160213/f501d630/attachment-0001.html>

More information about the OpenStack-dev mailing list