[
https://issues.apache.org/jira/browse/HDDS-1205?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16795814#comment-16795814
]
Yiqun Lin edited comment on HDDS-1205 at 3/19/19 8:28 AM:
----------------------------------------------------------
Just reviewed based on the v02 patch, some review comments from me:
1.
{code:java}
/*
* We don't take any action if the container is in OPEN state.
*/
if (state == LifeCycleState.OPEN) {
return;
}
{code}
Here we don't need to take action for some other state, like {{DELETED}} and
{{DELETING}}? I see these state containers were not be processed actually.
2. Can we add some debug log for send command details (which container, which
datanodes, which command behaviour) in
{{handleUnderReplicatedContainer/handleOverReplicatedContainer/handleInconsistentContainer}}?
It will be helpful for the trouble shooting.
3. Looks like old ReplicationManager hasn't been replaced by new
ReplicationManager. This will be done in another JIRA?
was (Author: linyiqun):
Just reviewed based on the v02 patch, some review comments from me:
1.
{code:java}
/*
* We don't take any action if the container is in OPEN state.
*/
if (state == LifeCycleState.OPEN) {
return;
}
{code}
Here we don't need to take action for some other state, like {{DELETED}} and
{{DELETING}}? I see these state containers were not be processed actually.
2. Can we add some debug log for send command details in
{{handleUnderReplicatedContainer/handleOverReplicatedContainer/handleInconsistentContainer}}?
It will be helpful for the trouble shooting.
3. Looks like old ReplicationManager hasn't been replaced by new
ReplicationManager. This will be done in another JIRA?
> Refactor ReplicationManager to handle QUASI_CLOSED containers
> -------------------------------------------------------------
>
> Key: HDDS-1205
> URL: https://issues.apache.org/jira/browse/HDDS-1205
> Project: Hadoop Distributed Data Store
> Issue Type: Improvement
> Components: SCM
> Reporter: Nanda kumar
> Assignee: Nanda kumar
> Priority: Major
> Labels: pull-request-available
> Attachments: HDDS-1205.000.patch, HDDS-1205.001.patch,
> HDDS-1205.002.patch
>
> Time Spent: 3h 10m
> Remaining Estimate: 0h
>
> This Jira is for refactoring the ReplicationManager code to handle all the
> scenarios that are possible with the introduction of QUASI_CLOSED state of a
> container.
> The new ReplicationManager will go through the complete set of containers in
> SCM to find out under/over replicated and unhealthy containers and takes
> appropriate action.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
