[
https://issues.apache.org/jira/browse/HDDS-5712?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jackson Yao updated HDDS-5712:
------------------------------
Description:
for now, when a balance iteration ended, we have to wait
`balancingInterval`(default 1 hour , too long) before start the next iteration.
since we already have the following code for waiting for each move option, the
only role of the waiting before starting a new iteration is to see whether the
balance thread has been interrupt.
{code:java}
ReplicationManager.MoveResult result = future.get(
config.getMoveTimeout().toMillis(), TimeUnit.MILLISECONDS);{code}
we can use isInterrupted to see whether it has been Interrupted, and thus we
can start the next iteration as soon as possible
was:when moving containers, we care about the replica changes very much. for
now , we can get replica infos only from the container report , which will be
sent every 60 minutes by default. this take a too long time. this Jira aims to
add a replicating container command ACK, from which scm can get the replica
changes as soon as possible
> no need to wait before start a new balance iteration
> ----------------------------------------------------
>
> Key: HDDS-5712
> URL: https://issues.apache.org/jira/browse/HDDS-5712
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Jackson Yao
> Priority: Major
>
> for now, when a balance iteration ended, we have to wait
> `balancingInterval`(default 1 hour , too long) before start the next
> iteration.
> since we already have the following code for waiting for each move option,
> the only role of the waiting before starting a new iteration is to see
> whether the balance thread has been interrupt.
> {code:java}
> ReplicationManager.MoveResult result = future.get(
> config.getMoveTimeout().toMillis(), TimeUnit.MILLISECONDS);{code}
> we can use isInterrupted to see whether it has been Interrupted, and thus we
> can start the next iteration as soon as possible
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
