[
https://issues.apache.org/jira/browse/HDFS-6788?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14081708#comment-14081708
]
Yongjun Zhang commented on HDFS-6788:
-------------------------------------
I thought I made a mistake in earlier patch after seeing Arpit's comment, but
after careful reviewing, it looks fine to me.
Hi [~andrew.wang],
I uploaded a patch to address the second comments of yours. For the extra
newline in the imports section, it's added by eclipse to separate the imports
comp.google and org.apache into different sections , and I think it's not bad
and didn't change it. Would you please take a look at the new revision? Thanks.
Hi [~arpitagarwal], thanks for reviewing it earlier, if you have time, I will
appreciate that you look at it again.
BTW Andrew, I saw that both FSDirection and FSNameSystem's lock code have quite
some indirection (thus a bit more runtime) when making a call, say,
{code}
void readLock() {
this.dirLock.readLock().lock();
}
{code}
It go through several "." and each of them is an indirection. This can be
improved. For example, we can create a class member, e.g., {{mReadLock =
this.dirLock.readLock();}}, then modify the readLock() methods to
{code}
void readLock() {
mReadLock.lock();
}
{code}
I can create a jira to change both classes if you agree. Thanks.
> Improve synchronization in BPOfferService with read write lock
> --------------------------------------------------------------
>
> Key: HDFS-6788
> URL: https://issues.apache.org/jira/browse/HDFS-6788
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode
> Affects Versions: 2.5.0
> Reporter: Yongjun Zhang
> Assignee: Yongjun Zhang
> Attachments: HDFS-6788.001.patch, HDFS-6788.002.patch
>
>
> Threads in DN (DataXceiver, PacketResponder, Async disk worker etc) may block
> at BPOfferService.getBlockPoolId() when calling BPOfferService.checkBlock(),
> though they are just reading the same blockpool id. This is unnecessary
> overhead and may cause performance hit when many threads compete. Filing this
> jira to replace synchronized method with read write lock
> (ReentrantReadWriteLock).
--
This message was sent by Atlassian JIRA
(v6.2#6252)
