[
https://issues.apache.org/jira/browse/HDFS-1547?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12978051#action_12978051
]
Sanjay Radia commented on HDFS-1547:
------------------------------------
>when a node is being-decommissioned state, why do you propose that it reduces
>the frequency ..
The reduced frequency is not for the "being-decommissioned" state but perhaps
for the "decommisioned" state.
I think longer term (perhaps another jira) a NN returns to a DN a state:
* "active" - the DN is registered with NN as active
* "being-decommissioned" - the DN is registered and is being decommission. This
state will be useful to break pipelines so that decommission can continue and
finish in spite of long running slow pipeline.
* "decommisioned" - you have been successfully decommissioned.
The heartbeat and BR intervals for "active" and "decommisioning" are the same.
It is not clear if the "decommisioned" state will have active heartbeats
(perhaps at a slower interval) or any BRs. This is longer discussion (Suresh
and I have debated this for long and haven't come to an agreement yet). This
state change and handshake protocol change is probably best done in a separate
jira.
> Improve decommission mechanism
> ------------------------------
>
> Key: HDFS-1547
> URL: https://issues.apache.org/jira/browse/HDFS-1547
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: name-node
> Affects Versions: 0.23.0
> Reporter: Suresh Srinivas
> Assignee: Suresh Srinivas
> Fix For: 0.23.0
>
>
> Current decommission mechanism driven using exclude file has several issues.
> This bug proposes some changes in the mechanism for better manageability. See
> the proposal in the next comment for more details.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
