[
https://issues.apache.org/jira/browse/HDDS-700?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16839049#comment-16839049
]
Sammi Chen commented on HDDS-700:
---------------------------------
Thanks [~xyao] for the comments.
bq. Line 353: if the caller does not set the location in the case of no
topology defined, should we have a default location like/default-location or
/default-rack.
Currently I will use /default-rack as the default location to be compatible
with current HDFS implementation. In future, we can refine this to get the
default location from network topology class.
{quote}
SCMContainerPlacementRackAware.java
Line 77: should we consider redefine the interface to cover case where we want
include the affinity node in the first node selection as well? For example,
client inside the cluster may prefer to write first copy to local container if
possible.{quote}
Good point. Will add a new parameter "List<DatanodeDetails> favoredNodes" to
cover this case, and other cases.
{quote} Line 84: should we consider some fallback strategy in certain cases
where we allow allocation succeeds but
Recover later when the placement if possible later?{quote}
Reasonable. But I think this should be related with the mininal acceptable
replication factor. For example, if the replication is 3 and minimal
replication is 1, then choose datanode should not fail if the mininal
replication can be satisfied. Currently there is no minimal replication
property in Ozone, also I am not clear if Ratis will still work when the
pipeline includes less than 3 nodes. Do you have this information?
{quote}Line 101-122: some of the logic can be consolidated.{quote}
Need more hint here.
All other comments will be addressed in next patch.
> Support rack awared node placement policy based on network topology
> -------------------------------------------------------------------
>
> Key: HDDS-700
> URL: https://issues.apache.org/jira/browse/HDDS-700
> Project: Hadoop Distributed Data Store
> Issue Type: Sub-task
> Reporter: Xiaoyu Yao
> Assignee: Sammi Chen
> Priority: Major
> Attachments: HDDS-700.01.patch
>
>
> Implement a new container placement policy implementation based datanode's
> network topology. It follows the same rule as HDFS.
> By default with 3 replica, two replica will be on the same rack, the third
> replica and all the remaining replicas will be on different racks.
>
> {color:#808080} {color}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
