Hi everyone.
I'd like first to thank you again for your involvement, and I really hope
I'm not making both of you lose your time. As you said maybe the deadlocks
I'm experiencing are due to FishFarm or Shoal, but to be fair I must add a
precision: it's very probable too that it's *my* fault if it doesn't work,
although I can't figure out how at all :-)
At this time, the test case I wrote is something very rustic as I'm not
familiar with JUnit, but I think it points out the issue quite correctly.
Thread dumps created with *visualvm* are ready, I'm going to send them to
Michael right now hoping it will be useful.
Let's keep in touch in case of an evolution, I'd be pleased to share
results!
Wishing you a nice week end,
Mathias Chouet
2008/8/7 Shreedhar Ganapathy <Shreedhar.Ganapathy_at_sun.com>
> Hi Michael
> Hope you are doing well.
> Thanks for jumping in to help. Much appreciated.
>
> Hi Mathias
> I very much appreciate your trying out these experiments. Needless to say,
> we have never put Shoal through this level of intensive message sizes.
> On one level, I am happy to see data like this so feel free to contribute
> data and extracted tests if possible to either FishFarm or to Shoal or both.
> It would help us a lot to have such tests in the community repository so it
> can become part of a smoke test.
>
> As Michael Bien says using a tool like VisualVM might provide a lot of
> insights. Will await Michael's assessments of the thread dump. If you can
> share those here we can also take a look.
>
> Cheers
> Shreedhar
>
>
> Michael Bien wrote:
>
> Hello Mathias,
>
> great to hear that you have still not giving up ;-)
>
> Well it is not entirely impossible that its FishFarm fault :) (thats why i
> filed a bug against ff). The problem is just that I can't reproduce it here,
> all my prior load tests with small messages where pretty reliable (5 out of
> 40k tasks where lost in action but recovered by fishfarm). The other problem
> is that I can't run the load tests again because I can't access the test
> cluster (holidays on my university...).
>
> could you next time it happens create a thread dump of the master node and
> just to be sure of the worker too (visualvm is a great tool for this
> purpose) and send it to me? Just to exclude deadlocks somewhere in FishFarm.
>
> Also if it is possible in some way to extract your *RasterAdapter *to a
> testcase this would be great.
>
> regards,
>
> Michael
>
> btw I am the FishFarm author
>
>