[
https://issues.apache.org/jira/browse/HDFS-1979?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13064800#comment-13064800
]
Eli Collins commented on HDFS-1979:
-----------------------------------
This is really nice. I like how BackupNodeProtocol is explicit now and we can
get rid of the spool file. The new tests look good.
Should setStorageInfo also setBlockPoolID and setClusterID? Eg in
BN#registerWith where we call all three I'm wondering why those aren't a single
method.
In FSIMage#getLastAppliedTxId why return editLog.getLastWrittenTxId()? It seems
like this method is for asking for the last tx from loading the image/edits log
(a static property), if someone wants the the last tx written to the edits log
(dynamic) they should use that API directly.
reloadFromImageFile is addressed in HDFS-2133, will comment there.
Why does LogLoadPlan() call super()?
Nits:
* In rollForwardByApplyingLog brace is on it's own line
* In downloadEditsToStorage pull f onto prev line
* BackupNodeProtocol isAlive and startLogSegment need javadocs
* FSNamesystem#registerBackupNode params could use a comment/better names to
distinguish which registration is which (eg bnReg, nnReg)
* BN "Is journal spooling in progress" comment should be replaced with a new
description. Eg the state transition: DROP_UNTIL_NEXT_ROLL --> JOURNAL_ONLY
<--> IN_SYNC
> HDFS-1073: Fix backupnode for new edits/image layout
> ----------------------------------------------------
>
> Key: HDFS-1979
> URL: https://issues.apache.org/jira/browse/HDFS-1979
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1979-prelim.txt, hdfs-1979.txt, hdfs-1979.txt
>
>
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
