[
https://issues.apache.org/jira/browse/HDDS-1348?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16813042#comment-16813042
]
Jitendra Nath Pandey commented on HDDS-1348:
--------------------------------------------
A few comments:
# Why don't we always call adjustBuffers on
{{xceiverClient.getReplicatedMinCommitIndex()}}? Isn't
replicated-min-commit-index is the source of truth in all cases.
# I think {{totalAckDataLength}} should be updated as soon as we get an ack,
instead of adjustBuffer. We are coupling the buffer-management with tracking of
acks.
# {{watchOnCommitIndex(boolean first)}}: It would be better to have two
different functions, one for buffer full case, and another for flush. A boolean
flag can be confusing for example, suppose when we call flush, buffer is full
as well, so it becomes tricky to understand code.
# Please add unit tests only for CommitWatcher class. The test should mock
xceiverClient, and test buffers are appropriately released, for exception as
well as non-exception cases.
# Let's make \{{CommitWatcher#getCommitIndex2flushedDataMap}} visible for
testing.
> Refactor BlockOutpuStream Class
> -------------------------------
>
> Key: HDDS-1348
> URL: https://issues.apache.org/jira/browse/HDDS-1348
> Project: Hadoop Distributed Data Store
> Issue Type: Improvement
> Components: Ozone Client
> Reporter: Shashikant Banerjee
> Assignee: Shashikant Banerjee
> Priority: Major
> Fix For: 0.5.0
>
> Attachments: HDDS-1348.000.patch
>
>
> BlockOutputStream contains functionalities for handling write, flush and
> close as well as tracking commitIndexes . The idea is to separate all
> commitIndex tracking and management code outside of BlockOutputStream class
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
