streams-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Matthew Hager [W2O Digital]" <mha...@w2odigital.com>
Subject Re: Bug in StreamsPersistWriterTask and StreamsProcessorTask. BREAKS LOCAL RUNTIME
Date Sat, 17 May 2014 01:43:20 GMT
If it pleases the community,

I have actually refactored a large portion of this for readability and
added a series of tests to ensure that it performs under different
conditions. I should have a PR for your review shortly.

Thanks!
Matthew



On 5/15/14, 10:19 AM, "Ryan Ebanks" <ryanebanks@gmail.com> wrote:

>Actually it should not be, !this.inQueue.isEmpty().  It should be, datum
>!=
>null
>
>
>On Wed, May 14, 2014 at 5:19 PM, Ryan Ebanks <ryanebanks@gmail.com> wrote:
>
>> I can't seem to log into Jira at the moment create the bug ticket.  But
>> this bug breaks the local runtime.
>>
>> A recent commit seems to have introduced this.  Its a simple fix, but
>>the
>> while statement on line 70 of StreamsProviderTask should read
>> while(this.keepRunning.get() || !this.inQueue.isEmpty()) {
>>
>> And line 75 in StreamsPersist writer task should read
>> while(this.keepRunning.get() || !this.inQueue.isEmpty()) {
>>
>>
>> Both lines are missing the !this.inQueue.isEmpty() statement.  This
>>causes
>> tasks to be shutdown before data is processed and data is orphaned in
>>the
>> stream.  In the worst case only the provider runs.
>>
>> -Ryan Ebanks


Mime
View raw message