Orbited bugs

35 views
Skip to first unread message

P.R.

unread,
May 25, 2011, 11:35:58 AM5/25/11
to orbite...@googlegroups.com

Hello,
 
I run Orbited (using script from this article http://cometdaily.com/2008/10/10/scalable-real-time-web-architecture-part-2-a-live-graph-with-orbited-morbidq-and-jsio/ ) and picked up some bugs:

 - JS scripts don't work under latest Opera and has some strange behaviour under IE 7 and Chrome (is there any jQuery interface?)
 - memory usage of python process was raising all the time, >= 100 Mb i killed process (any idea how to solve this memory lake?)
 
Is there any developer version (official site is still broken ;/)?


Thanks.

Jared Wilson

unread,
May 25, 2011, 11:42:05 AM5/25/11
to orbite...@googlegroups.com
I'm not sure if still support for orbited&morbid. Most development has been moved towards http://hookbox.org
Might want to look into using that.

Note: With my experience using orbited & morbid if you use setuptools "easy_install" to install orbited you get all the dependencies installed automatically.

Hope this helps!

--
You received this message because you are subscribed to the
Orbited discussion group.
To post, send email to
<orbite...@googlegroups.com>
To unsubscribe, send email to
<orbited-user...@googlegroups.com>
For more options, visit
<http://groups.google.com/group/orbited-users>

P.R.

unread,
May 25, 2011, 12:01:07 PM5/25/11
to orbite...@googlegroups.com
Thanks for answer.

I'm surprised that orbited&morbid is no longer under development, hmm.

Heh, it's some doom or something that hookbox.org domain is broken too? :-)

Regards.

Niklas B

unread,
May 26, 2011, 5:29:48 AM5/26/11
to Orbited Discussion
It's mostly due to time constraints, but me for example plan on keep
running Orbited for a long time. Develop too.

It's a memory leak, should be fixed in the SVN/Mercury rep

P.R.

unread,
May 26, 2011, 6:26:49 AM5/26/11
to orbite...@googlegroups.com
Thank you Nikolas.
 
I run orbited from https://bitbucket.org/desmaj/orbited (orbited-0.7.11beta3.tar.gz) ... is there any API documentation, how to write producer? My old producer script could not connect to the source:
 
Connection failed. Reason: [Failure instance: Traceback (failure with no frames)
: <class 'twisted.internet.error.ConnectionRefusedError'>: Connection was refuse
d by other side: 10061: No connection could be made because the target machine a
ctively refused it..
]
 
There is some DeprecationWarning in orbited (I'm using Twisted 11.0.0)
 
DeprecationWarning: listenWith is deprecated since Twisted 10.1.  See IReactorFDSet.
  reactor.listenWith(port_class, factory=factory_class(), resource=root, childName=child_path)
 
 
Thanks!
 
 
 

P.R.

unread,
May 26, 2011, 6:45:12 AM5/26/11
to orbite...@googlegroups.com
Hmm, it does not resolve problems, there are still bugs in Opera and memory lake ;///

Niklas B

unread,
May 26, 2011, 7:20:16 AM5/26/11
to Orbited Discussion
That's not good at all, I'm using a modified version of 0.8 so I'm not
sure how to solve it best.

Anyone else remember how the memory leak was fixed?

P.R.

unread,
May 26, 2011, 7:38:06 AM5/26/11
to orbite...@googlegroups.com
Hmm, where can I find 0.8 branch ? http://orbited.org/svn does not work.

Niklas B

unread,
May 26, 2011, 7:48:32 AM5/26/11
to Orbited Discussion
Afraid it wouldn't help you, the bug is that (not sure if it's only
when connected to stomp) it doesn't properly close the connections.
What I did was a workaround:

1. Each of my clients connect to STOMP + IRC
2. When IRC closes the connection (since IRC has built in timeout)
Orbited also closes STOMP manually

Aa`Koshh

unread,
May 27, 2011, 9:14:36 AM5/27/11
to Orbited Discussion
Hi,

I doubt that it could cause a memory leak of such proportions, but I
think that in MorbidQ topics are never destroyed.
There is a function for it in the MessageQueueManager, but I don't
think it ever gets called, especially since there is a syntax error in
remove_pubsub_child( missing a "self." before message_queues), at
least in the version I use.
Because of this error I had a growing number of disconnected stomp
protocols lingering in the morbid topic queues, but that was only
after I added the conditional call to destroy_queue in
unsubscribe_all_queues. Don't know if the memory leak is gone finally,
I will keep checking it with manhole.

Bests

Reply all
Reply to author
Forward
0 new messages