hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Anu Engineer (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-11184) Ozone: SCM: Make SCM use container protocol
Date Wed, 22 Feb 2017 23:33:44 GMT

    [ https://issues.apache.org/jira/browse/HDFS-11184?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15879470#comment-15879470

Anu Engineer commented on HDFS-11184:

Thanks for a very detailed code review, [~xyao].  I have addressed all comments in patch 6.
Please see the specifics below.

bq. Line 95: NIT: "HDFS" can be removed

bq. Line 100: May need to update with the new SCM protocol.
Not sure I understood this comment clearly, Do you think I should add more info in the comment

bq. Line 103: @InterfaceAudience.Private should we use the LimitedPrivate here as it will
likely for both HDFS and HBase
Fixed, added HDFS, HBASE, OZONE and CBLOCK as the likely clients.

bq. Line 122,123: NIT: suggest rename scmRpcServer -> clientRpcServer scmRpcAddress->clientRpcAddress
bq. Line 139: Do we have a ticket for fixing the SCM ClusterID generation?
Thanks for catching this, Just filed HDFS-11442

bq. Line 159: should we move the log message to StorageContainerManager#start() where the
rpc server actually started?
Fixed, moved both RPC server messages to start.

bq. Line 186/206/235/272/288: remove static
Leaving it  same, based on an off-line chat with [~xyao]

bq. Line 190: the string concat for host and port is not necessary? Can we use addr and its
toString() implicitly
Good catch, fixed.

bq. Line 241: should we update the host part when updateListenAddress() if the host is allowed
to be configured as
We are just changing the config in memory, so not sure what is the right thing to do here.
Please let me know your suggestions.

bq. Line 298: let's open a separate ticket for SCM HA support if there isn't one yet.
 Done, HDFS-11443

bq. Line 357: can we add a log message for the rpc server stop?

bq. Line 91: NIT: typo contianer -> container

bq. Line 88: javadoc does not match the actual code, but we can fix that later.
Added a TODO to reflect this in code.

bq. Line 230: javadoc should be the client that perform the call
Thanks, fixed this error.

bq. Line 340: can we open a ticket to fix the datanode id issue in the TODO?
Done, HDFS-11444

bq. Line 57: if (cluster != null)
Thanks, fixed.

> Ozone: SCM: Make SCM use container protocol
> -------------------------------------------
>                 Key: HDFS-11184
>                 URL: https://issues.apache.org/jira/browse/HDFS-11184
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: ozone
>    Affects Versions: HDFS-7240
>            Reporter: Anu Engineer
>            Assignee: Anu Engineer
>             Fix For: HDFS-7240
>         Attachments: HDFS-11184-HDFS-7240.001.patch, HDFS-11184-HDFS-7240.002.patch,
HDFS-11184-HDFS-7240.003.patch, HDFS-11184-HDFS-7240.004.patch, HDFS-11184-HDFS-7240.005.patch,
> SCM will start using container protocol to communicate with datanodes. 
> This change introduces some test failures due to some missing features which will be
moved to KSM. Will file separate JIRA to track disabled ozone tests. 

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org

View raw message