[
https://issues.apache.org/jira/browse/HBASE-14268?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14731049#comment-14731049
]
stack commented on HBASE-14268:
-------------------------------
To be clear, my question is if weak references require a 'Full GC' before they
are cleared. I do not think this the case (We use weak references elsewhere in
the code base so it must be OK). I think it soft references that require a Full
GC but it is a while since I messed with this stuff and my recollection is
dodgy. I googled around some but could not find a definitive answer. My fear is
that if it requires a 'Full GC', then we will accumulate weak references for
ever given users do all in their power to put off a 'Full GC' since 'Full GC'
is catastrophic.
Agree that continuously getting a lock is not normal usage and that continuous
adding/removing may overrun the purging runs leading to filled heaps or Full
GC. If weak references are removed without a Full GC, and given that usage is
low of locks relatively, we should be good (add a log every time something is
removed from the queue while you have GC logging enabled? Make sure you don't
have to wait on a Full GC to get a log?)
bq. ...That is because the range under the locks should be as small as possible.
bq. ....but I think using array is simpler and faster than using TreeSet or
Collections.sort(List).
Makes sense. Thank you.
> Improve KeyLocker
> -----------------
>
> Key: HBASE-14268
> URL: https://issues.apache.org/jira/browse/HBASE-14268
> Project: HBase
> Issue Type: Improvement
> Components: util
> Reporter: Hiroshi Ikeda
> Assignee: Hiroshi Ikeda
> Priority: Minor
> Fix For: 2.0.0, 1.3.0
>
> Attachments: 14268-V5.patch, HBASE-14268-V2.patch,
> HBASE-14268-V3.patch, HBASE-14268-V4.patch, HBASE-14268-V5.patch,
> HBASE-14268-V5.patch, HBASE-14268-V6.patch, HBASE-14268-V7.patch,
> HBASE-14268-V7.patch, HBASE-14268.patch, KeyLockerIncrKeysPerformance.java,
> KeyLockerPerformance.java
>
>
> 1. In the implementation of {{KeyLocker}} it uses atomic variables inside a
> synchronized block, which doesn't make sense. Moreover, logic inside the
> synchronized block is not trivial so that it makes less performance in heavy
> multi-threaded environment.
> 2. {{KeyLocker}} gives an instance of {{RentrantLock}} which is already
> locked, but it doesn't follow the contract of {{ReentrantLock}} because you
> are not allowed to freely invoke lock/unlock methods under that contract.
> That introduces a potential risk; Whenever you see a variable of the type
> {{RentrantLock}}, you should pay attention to what the included instance is
> coming from.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
