RFR: 7879: Automated Analysis taking very long time to produce results for Class Leak Rule and showing wrong results. [v6]

Brice Dutheil bdutheil at openjdk.org
Tue Jan 17 15:15:41 UTC 2023


On Sun, 8 Jan 2023 19:49:01 GMT, Suchita Chaturvedi <schaturvedi at openjdk.org> wrote:

>> Suchita Chaturvedi has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Removed the synchronized block and used ConcurrentLinkedQueue instead
>
> In the latest patch, I have removed the synchronized block which was the main bottle neck for the approval of this PR. Also, used the suggested data structure i.e. ConcurrentLinkedQueue for the processing of records. 
> 
> I have debugged in detail and figured out that the reason for inconsistent results for the customer JFR which has around 65000 class entries to process was the time multithreaded tasks were taking to complete which was dependent on many factors like how many cores our CPU has OR how busy our CPU is when its performing those tasks. Hence, for safer side I have added a check to see whether all tasks are performed or not. If not, I have added a console message to alert the user for increasing the configured timeout. The default timeout is 5 min which is quite sufficient to complete all the tasks still if the count of entries are too much and there are some pending tasks, user will be alerted for the same.
> 
> @bric3 Please let me know when can we have a zoom (if required).

Hi @Suchitainf I'll have a look tommorrow.

-------------

PR: https://git.openjdk.org/jmc/pull/419


More information about the jmc-dev mailing list