[
https://issues.apache.org/jira/browse/PROTON-1497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16054013#comment-16054013
]
Jiri Danek commented on PROTON-1497:
------------------------------------
bq. The problem is that if the calculation takes a long time (>1min) to finish,
randomly the program crashes when trying to execute the send instruction (line
157).
As a workaround, I'd use two threads. One runs the reactor, the other
calculates the hashes. I would use a
https://docs.python.org/2/library/queue.html to pass messages from the reactor
thread to the processing thread, and another one for the other direction. I
would use
https://qpid.apache.org/releases/qpid-proton-0.17.0/proton/python/api/proton.reactor.EventInjector-class.html
to tell the reactor that there is something to send.
Reading previous paragraph, it sounds absurd to use queues in Python when I
already have a broker to provide queuing. IMO, there just shouldn't be timeout
if I take my time in the on_message method. There should be some background
thread in Proton Python that would keep the session alive for me, without me
worrying about it.
bq. I also tried to encapsulate the send into a try/except clause, but the code
inside the except never gets executed.
That makes sense, IMO. Proton Python reactor is event-driven, so exceptions
won't work. The errors need to be handled in reactor error handler method(s).
{{on_connection_error}} and so on.
> python program crash on send msg to queue
> -----------------------------------------
>
> Key: PROTON-1497
> URL: https://issues.apache.org/jira/browse/PROTON-1497
> Project: Qpid Proton
> Issue Type: Bug
> Components: proton-c, python-binding
> Affects Versions: 0.17.0
> Environment: redhat 7.2 - activemq-5.14.3 - Python 3.6.0
> Reporter: Pier Paolo Panto
> Attachments: md5.py
>
>
> In the attached file there is an on_message method that performs a lenghty
> calculation whenever a message is receivede on the input queue. In this case,
> the branch that gets executed is the one starting at line 123.
> A the end of the branch a new message is created and sent so an out queue.
> The problem is that if the calculation takes a long time (>1min) to finish,
> randomly the program crashes when trying to execute the send instruction
> (line 157). I also tryed to encapsulate the send into a try/except clause,
> but the code inside the except never gets executed.
> Moreover, the exit code (echo $?) when the program crashes is 0
> When this happens, I get this message in the log file:
> handlers.py:234 in print_error(): local-idle-timeout expired
> but setting the timeout parameters for the connection has no effect.
> What could be causing this behaviour?
> Best Regard,
> PPP
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org
