Database access from bg thread: python

[Rod Fitzsimmons Frey]

Hi. I posted yesterday but I don't think I gave enough information.

I'm getting a crash when I try to create an embedded database from
python from within the consumer thread in a producer-consumer setup.
I'm sure I'm abusing the interface somehow, but I don't know how and
I'm stuck.

Here's a sample program that manifests the crash:

https://gist.github.com/1958654

Here's the dump that results:

https://gist.github.com/1958707

The same error occurs if I try to create a db instance in the
foreground thread as a global and access it from the bg thread.

Thanks for any advice/insight/help!

Some other info:

rfrey@oilers:~/dev/rowzero$ java -version
java version "1.7.0_03"
Java(TM) SE Runtime Environment (build 1.7.0_03-b04)
Java HotSpot(TM) Client VM (build 22.1-b02, mixed mode)

rfrey@oilers:~/dev/rowzero$ python -V
Python 2.6.5

Using neo4j-community-1.6

[Peter Neubauer]

Rod,
seems you are running java7 there? Have no idea about Python, but
maybe you could switch JRE to Java6 and try again? Seems the JVM is
SegFaulting ...

Cheers,

/peter neubauer

G:  neubauer.peter
S:  peter.neubauer
P:  +46 704 106975
L:   http://www.linkedin.com/in/neubauer
T:   @peterneubauer

Neo4j 1.6 released                 - dzone.com/6S4K
The Neo4j Heroku Challenge   - http://neo4j-challenge.herokuapp.com/

[Jacob Hansson]

Hmm.. looks like JPype segfaults. Would have to dig into it with a debugger I think to figure out what is going on, but I'm gonna venture a guess that it is a bug.

We do test running neo4j with threads on python as part of the build though. It might help you to look at how those are set up:

https://github.com/neo4j/python-embedded/blob/master/src/test/python/threads.py

Let me know if that does not work!

--
Jacob Hansson
Phone: +46 (0) 763503395
Twitter: @jakewins

[Rod Fitzsimmons Frey]

Hey guys, thanks for the kind attention. I kind of left this behind
by moving over to a server instance of Neo4j and using the
neo4jrestclient python module. Solved the problem but might not be
performant enough in the future.

I agree it's a jpype problem: I just reproduced using the vanilla
jpype library, no neo4j involved. Could it have to do with the
jpype.attachThreadToJVM() call that one is supposed to execute when in
a background thread? If so, should that be the client's
responsibility, or the neo4j library's?


On Mar 15, 4:42 am, Jacob Hansson <jacob.hans...@neotechnology.com>
wrote:

> Hmm.. looks like JPype segfaults. Would have to dig into it with a debugger
> I think to figure out what is going on, but I'm gonna venture a guess that
> it is a bug.
>
> We do test running neo4j with threads on python as part of the build
> though. It might help you to look at how those are set up:
>

> https://github.com/neo4j/python-embedded/blob/master/src/test/python/...

>
> Let me know if that does not work!
>
> On Wed, Mar 14, 2012 at 10:40 PM, Peter Neubauer <
>
>
>
>
>
>
>
>
>
> peter.neuba...@neotechnology.com> wrote:
> > Rod,
> > seems you are running java7 there? Have no idea about Python, but
> > maybe you could switch JRE to Java6 and try again? Seems the JVM is
> > SegFaulting ...
>
> > Cheers,
>
> > /peter neubauer
>
> > G:  neubauer.peter
> > S:  peter.neubauer
> > P:  +46 704 106975
> > L:  http://www.linkedin.com/in/neubauer
> > T:   @peterneubauer
>
> > Neo4j 1.6 released                 - dzone.com/6S4K

> > The Neo4j Heroku Challenge   -http://neo4j-challenge.herokuapp.com/
>
> > On Fri, Mar 2, 2012 at 3:31 PM, Rod Fitzsimmons Frey <rodf...@gmail.com>