hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Flavio Paiva Junqueira (JIRA)" <j...@apache.org>
Subject [jira] Issue Comment Edited: (HADOOP-5188) Modifications to enable multiple types of logging
Date Sun, 31 May 2009 21:36:07 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-5188?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714929#action_12714929
] 

Flavio Paiva Junqueira edited comment on HADOOP-5188 at 5/31/09 2:35 PM:
-------------------------------------------------------------------------

Luca, I was wondering if you could clarify a few points of your last post:

# I'm not sure I follow the argument about append. More specifically, I'm not sure if your
argument is about implementation complexity or just abstraction. To implement append with
BookKeeper, can't we simply open another ledger upon a call to append? With respect to abstraction,
it sounds rights that it would be better not to have to implement "append" for bookkeeper
logging because no such operation is supported directly by the service.
# On your argument about not fully using LogDevice, perhaps you want to elaborate on it a
little more. Why is it ok to leave it partially implemented?  I'm specially concerned about
someone looking at this code later and having a hard time understanding it because there is
a mix of mechanisms implemented. If the remainder of the abstraction can be implemented right
away, then it might be ok to postpone in the interest of  progress.
# I didn't follow the correlation between the BackupNode and logging with BookKeeper. Are
you saying that if we don't have LogDevice, then we would have to spawn a new process to perform
BookKeeper logging as with the BackupNode? Independent of the answer to this question, please
clarify.
  

      was (Author: fpj):
    Luca, I was wondering if you could clarify a few points of your last post:

# I'm not sure I follow the argument about append. More specifically, I'm not sure if your
argument is about implementation complexity or just abstraction. To implement append with
BookKeeper, can't we simply open another ledger upon a call to append? With respect to abstraction,
it sounds rights that it would be better not to have to implement "append" for bookkeeper
logging because no such operation is supported directly by the service.

# On your argument about not fully using LogDevice, perhaps you want to elaborate on it a
little more. Why is it ok to leave it partially implemented?  I'm specially concerned about
someone looking at this code later and having a hard time understanding it because there is
a mix of mechanisms implemented. If the remainder of the abstraction can be implemented right
away, then it might be ok to postpone in the interest of  progress.

# I didn't follow the correlation between the BackupNode and logging with BookKeeper. Are
you saying that if we don't have LogDevice, then we would have to spawn a new process to perform
BookKeeper logging as with the BackupNode? Independent of the answer to this question, please
clarify.
  
  
> Modifications to enable multiple types of logging 
> --------------------------------------------------
>
>                 Key: HADOOP-5188
>                 URL: https://issues.apache.org/jira/browse/HADOOP-5188
>             Project: Hadoop Core
>          Issue Type: Improvement
>          Components: dfs
>    Affects Versions: 0.20.0
>            Reporter: Luca Telloli
>             Fix For: 0.21.0
>
>         Attachments: HADOOP-5188.patch, HADOOP-5188.patch, HADOOP-5188.patch, HADOOP-5188.patch,
HADOOP-5188.patch, HADOOP-5188.patch, HADOOP-5188.pdf
>
>


-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message