[
https://issues.apache.org/jira/browse/MYRIAD-11?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Santosh Marella updated MYRIAD-11:
----------------------------------
Description:
- Implement more scale down policies (currently we have "NMs with least AMs"
policy). Some ideas:
- Kill NMs with least number of MapReduce AMs (because MR has HA)
- Kill NMs with "most recently launched" AMs.
- Kill NMs with "certain labels" (With Hadoop v2.6 YARN supports Label
Based Scheduling)
- Kill enough NMs to free up "X cpus, Y mem"
was:
- Implement more scale down policies (currently we have "NMs with least AMs"
policy). Some ideas:
- Kill NMs with least number of MapReduce AMs (because MR has HA)
- Kill NMs with "most recently launched" AMs.
- Kill NMs with "certain labels" (With Hadoop v2.6 YARN supports Label
Based Scheduling)
> Implement more "NM scale down" policies
> ---------------------------------------
>
> Key: MYRIAD-11
> URL: https://issues.apache.org/jira/browse/MYRIAD-11
> Project: Myriad
> Issue Type: Improvement
> Reporter: Santosh Marella
>
> - Implement more scale down policies (currently we have "NMs with least AMs"
> policy). Some ideas:
> - Kill NMs with least number of MapReduce AMs (because MR has HA)
> - Kill NMs with "most recently launched" AMs.
> - Kill NMs with "certain labels" (With Hadoop v2.6 YARN supports Label
> Based Scheduling)
> - Kill enough NMs to free up "X cpus, Y mem"
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
