[ 
https://issues.apache.org/jira/browse/HDDS-4560?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Glen Geng updated HDDS-4560:
----------------------------
    Description: 
Add ReadWriteLock into PipelineStateMap to protect contentions between 
RaftServer and PipelineManager.

If current leader SCM (as leader of term N) steps down while there are on-going 
transactions, there might be read/write contentions that new leader SCM (as 
leader of term N+1) has replicated some raft log entries encapsulating method 
addPipeline/removePipeline/updatePipelineState to the underlying RaftServer 
that are being applied by StateMachineUpdater, and at the same time 
PipelineManager is calling getPipelines.

  was:
Add ReadWriteLock into PipelineStateMap to protect contentions between 
RaftServer and PipelineManager.

If current leader SCM (as leader of term N) steps down while there are on-going 
transactions, there might be read/write contentions that new leader SCM (as 
leader of term N+1) has replicated some raft log entries encapsulating method 
addPipeline/removePipeline/updatePipelineState to underlying RaftServer that 
are being applied by StateMachineUpdater, and at the same time PipelineManager 
is calling getPipelines.


> Add ReadWriteLock into PipelineStateMap to protect contentions between 
> RaftServer and PipelineManager
> -----------------------------------------------------------------------------------------------------
>
>                 Key: HDDS-4560
>                 URL: https://issues.apache.org/jira/browse/HDDS-4560
>             Project: Hadoop Distributed Data Store
>          Issue Type: Sub-task
>          Components: SCM HA
>    Affects Versions: 1.1.0
>            Reporter: Glen Geng
>            Assignee: Glen Geng
>            Priority: Major
>
> Add ReadWriteLock into PipelineStateMap to protect contentions between 
> RaftServer and PipelineManager.
> If current leader SCM (as leader of term N) steps down while there are 
> on-going transactions, there might be read/write contentions that new leader 
> SCM (as leader of term N+1) has replicated some raft log entries 
> encapsulating method addPipeline/removePipeline/updatePipelineState to the 
> underlying RaftServer that are being applied by StateMachineUpdater, and at 
> the same time PipelineManager is calling getPipelines.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to