Github user jvrao commented on the issue:

    https://github.com/apache/bookkeeper/pull/189
  
    Right. So if there are two dirs /journal1 and /journal2 both were created
    on the same disk.
    With the dedup, we will choose only one say, /journal1 and /journal2 is
    always kept empty. Am I getting it right?
    We do the dedup logic even while getting the disk capacity etc etc. So here
    we are masking/managing mis=configuration
    
    But our point is, configuring two directories under same disk partition is
    itself wrong, unless user wanted to do it explicitly
    for some test reasons. That is why we have that configuration variable,
    that forces operator error, unless it is set explicitly " I know what I am
    doing" kind.
    
    Makes sense?
    
    On Tue, Jun 27, 2017 at 2:10 PM, Sijie Guo <notificati...@github.com> wrote:
    
    > @jvrao <https://github.com/jvrao>
    >
    > sorry, what is the performance problem behind this? is there anything I
    > missed in this conversation?
    >
    > If a user configures multiple journal directories, can we dedup and use
    > one journal directory per disk partition? If we dedup, there is no
    > performance degradation and there is no backward compatibility issue, 
right?
    >
    > For example, we open a entrylog file by calling
    > *ledgerDirsManager.getWritableLedgerDirsForNewLog()* , the dirs manager
    > can handle the duplication (make sure one directory per disk partition is
    > used) and return the right directory. By doing this, all the 
directory/disk
    > partition related complexity is hidden and managed and people don't have
    > the know the details.
    >
    > I don't have any strong opinion adding another setting. My feeling is that
    > we are not taking the right approach to address the problem here.
    >
    > —
    > You are receiving this because you were mentioned.
    > Reply to this email directly, view it on GitHub
    > <https://github.com/apache/bookkeeper/pull/189#issuecomment-311486531>,
    > or mute the thread
    > 
<https://github.com/notifications/unsubscribe-auth/AAChrujqc3KnxxpUHY4AZ82vXNU7zblzks5sIW_EgaJpZM4N4EFC>
    > .
    >
    
    
    
    -- 
    Jvrao
    ---
    First they ignore you, then they laugh at you, then they fight you, then
    you win. - Mahatma Gandhi



---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

Reply via email to