[
https://issues.apache.org/jira/browse/HDDS-2?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16458998#comment-16458998
]
Bharat Viswanadham edited comment on HDDS-2 at 5/1/18 12:01 AM:
----------------------------------------------------------------
Hi [~anu]
Thank You for your comments.
To start replication manager, we can use chillModeStatus. When it is out of
chill mode, we can start replication manager.
And for readPath, when getContainerInfo is called, we just return containerInfo
from containerdb, so when client tries to acquire pipeline it will have
exception. (so, we don't check whether it is out of chill mode or not)
And for write-path also, we can remove chill mode check, and when we return a
pipeline matching with given details, and when it tries to acquire a pipeline
it will fail at the client end.
Or we can make some changes in SCM, to add map like you have proposed and
intersect it when allocating containers. But this will not change the chill
mode logic proposed in this jira.
Please let me know your comments and also let me know if I am missing something
here.
was (Author: bharatviswa):
Hi [~anu]
Thank You for your comments.
To start replication manager, we can use chillModeStatus. When it is out of
chill mode, we can start replication manager.
And for readPath, when getContainerInfo is called, we just return containerInfo
from containerdb, so when client tries to acquire pipeline it will have
exception. (so, we don't check whether it is out of chill mode or not)
And for write-path also, we can remove chill mode check, and when we return a
pipeline matching with given details, and when it tries to acquire a pipeline
it will fail at the client end. (This will be addressed in later jira)
Or we can make some changes in SCM, to add map like you have proposed and
intersect it when allocating containers. But this will not change the chill
mode logic proposed in this jira.
Please let me know your comments and also let me know if I am missing something
here.
> Chill Mode to consider percentage of container reports
> ------------------------------------------------------
>
> Key: HDDS-2
> URL: https://issues.apache.org/jira/browse/HDDS-2
> Project: Hadoop Distributed Data Store
> Issue Type: Task
> Components: SCM
> Reporter: Bharat Viswanadham
> Assignee: Bharat Viswanadham
> Priority: Major
> Fix For: 0.2.1
>
> Attachments: Chill Mode.pdf, HDDS-02.002.patch, HDFS-13500.00.patch,
> HDFS-13500.01.patch, HDFS-13500.02.patch
>
>
> To come out of chill mode currenly if one datanode is registered, we come out
> of chill mode in SCM.
> This needs to be changed to consider percentage of container reports.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
