Inlined. On Fri, Jun 19, 2015 at 4:17 AM, zhou weitao <zhouwtl...@gmail.com> wrote:
> Alex, hi, > > 2015-06-18 23:25 GMT+08:00 Alex Rukletsov <a...@mesosphere.com>: > >> Zhou, >> >> I haven't read the *Design* yet, but I don't think it is solving the same >>> question between priorities and quota. For example, assume we only have 10G >>> memory reservating for framework A totally, then another urgency framework >>> is getting nothing. which is statical partition still. While priority can >>> pre-empt that. >>> >> >> I'm not sure what is your concern here. If all you have is 10GB and both >> your frameworks A and B may need 10GB each at the same time, you definitely >> need to add more RAM : ). Mesos uses fair sharing for distributing >> resources among frameworks, quota will not be an exception. If all you have >> is 10GB and for both A and B you have reserved 10GB, total reserved >> resources are 20GB, which means your cluster is under quota. I would say, >> if this is happening in production cluster, several devops engineers should >> have already been paged : ). >> >> It's up to allocator implementation to decide what to do in this >> situation. An obvious approach is to throttle (i.e. revoke resources) both >> frameworks proportionally to their role weights. Quota does not introduce >> static partitioning, it rather guarantees, a production framework gets >> enough resources regardless of any events happening in the cluster, given >> these resources are available. >> > > I am sorry for my poor English firstly. In a nutshell, what I pointed is > that it is different between priorities and mesos quota. Here I am confused > about the following 2 question: > No worries, your English is good enough to understand what you mean. > 1. The *quota* will be always reserved for some framework in the cluster? > And other frameworks are all forbidden to use it? > The quota is being designed right now and the draft has not been published yet. Our current idea is to set quota per role and therefore leverage role weights. Other framework will be able to use resources from unused quotas, but will have to release the resources at the instance they are required by the framework, which "owns" them. > 2. I am figuring out such scenario, as the above Brian said, I have 2 > framework: Framework A is responsible for Real-time job, high priority but > low frequency of use, while framework B is responsible for offline job, > always running, greedy for resource, but low priority. Then, A and B are > using the same mesos cluster, how can I config to let B use all of the > resource till A pre-empt some of that? > An operator sets up quota for A, which is unused most of the time and B will be offered resources marked as revocable from A's quota. If A needs to launch more tasks, B tasks will be preempted. > > > thanks a lot. > > weitao Zhou > >
