Arun Suresh commented on YARN-2882:

Glad you could chime in [~kasha].

bq. Do the opportunistic containers in YARN-1011 and here share characteristics 
enough to be the same? ..
Although the QUEUEABLE containers will be allocated opportunistically, its 
naming has more to do with the fact that it will be queued at some NM. Given 
the fact that the latest patch introduces a new enum type (ExecutionType), 
instead extending the ContianerType, my vote is to probably add SPECULATIVE / 
OPPERTUNISTIC as a new ExecutionType for YARN-1011 in addition to QUEUEABLE to 
signal different mechanisms for handling the type of Request Ask.

bq. To set the cgroups and OOM priorities, we need the type in container 
initialization context...
Agreed.. But given that this patch is addressing mostly API changes, I'd prefer 
adding it in a subsequent patch when required.

bq. We are considering developing YARN-1011 also on a branch. Any ideas on how 
we can share code?
I like the Idea of pushing this to trunk, if we are limiting this to addition 
of the ExecutionType

> Add ExecutionType to denote if a container execution is GUARANTEED or 
> -------------------------------------------------------------------------------
>                 Key: YARN-2882
>                 URL: https://issues.apache.org/jira/browse/YARN-2882
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: nodemanager, resourcemanager
>            Reporter: Konstantinos Karanasos
>            Assignee: Konstantinos Karanasos
>         Attachments: YARN-2882-yarn-2877.001.patch, 
> YARN-2882-yarn-2877.002.patch, yarn-2882.patch
> This JIRA introduces the notion of container types.
> We propose two initial types of containers: guaranteed-start and queueable 
> containers.
> Guaranteed-start are the existing containers, which are allocated by the 
> central RM and are instantaneously started, once allocated.
> Queueable is a new type of container, which allows containers to be queued in 
> the NM, thus their execution may be arbitrarily delayed.

This message was sent by Atlassian JIRA

Reply via email to