<br><br><div><span class="gmail_quote">On 11/22/06, <b class="gmail_sendername">Brian Granger</b> <<a href="mailto:ellisonbg.net@gmail.com">ellisonbg.net@gmail.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Do you have it working now though?<br><br>On 11/22/06, Flavio Coelho <<a href="mailto:fccoelho@gmail.com">fccoelho@gmail.com</a>> wrote:<br>> Thanks Fernando,<br>><br>> I thinks this arrangement looks fine. The only change I can think of is to
<br>> handle the errors that end up in the logs in such a way that the user also<br>> gets informed of them.<br><br>I agree this would be nice, I think the best way is to simply report<br>*where* the log files are located and tell the user to look at them if
<br>there are problems. I don't think we want to display all the log<br>files by default though.</blockquote><div><br><br><br>definetly not all the logs but information about errors and to fix them. <br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
> In my example, if you hadn't pointed me to the logs I wouldn't have figured<br>> out what was going on. The ipcluster script should output a dignostic of the<br>> whole startup process, or at least spit out some meaningful error messages.
<br>> What do you think about it?<br><br><br><br>> Flávio<br>><br>><br>> On 11/21/06, Fernando Perez <<a href="mailto:fperez.net@gmail.com">fperez.net@gmail.com</a>> wrote:<br>> > On 11/21/06, Brian Granger <
<a href="mailto:ellisonbg.net@gmail.com">ellisonbg.net@gmail.com</a>> wrote:<br>> > > It looks like the engines are starting but are not finding the<br>> > > controller you have running. Things to check:
<br>> > ><br>> > > 1. Do you have a firewall blocking ports on the controller's computer?<br>> > ><br>> > > 2. Did you tell the engines where the controller is running:<br>> > >
<br>> > > ipengine --controller-ip=[ip of the machine running the controller]<br>> > ><br>> > > Another way of finding out what is going on is to start the<br>> > > ipcontroller and ipengine with the -l flags:
<br>> > ><br>> > > ipcontroller -l cont<br>> > > ipengine -l eng<br>> > ><br>> > > These will cause the log files to be written to disk in the format:<br>> > ><br>> > >
contpid.log<br>> > > engpig.log<br>> ><br>> > Also, note that if you are using ipcluster to start things, it makes<br>> > all logs go to<br>> ><br>> > ~/.ipython/ipcluster-PID1-{con|eng}-
HOST-PID2.log<br>> ><br>> > where:<br>> ><br>> > - PID1: PID of the starting process (the ipcluster script). This<br>> > simply allows you to have a common lexicographical sorting tag for all
<br>> > processes (controller and engines) coming from a given invocation of<br>> > iplucster.<br>> ><br>> > - 'con' or 'eng': tag to mark whether that log file is for a<br>> > controller or an engine
<br>> ><br>> > - HOST: host where that particular process is actually running.<br>> ><br>> > - PID2: PID of that particular process (controller or engine).<br>> ><br>> > This format seems to allow easy automatic organization, debugging and
<br>> > cleanup of logfiles, feel free to suggest improvements.<br>> ><br>> > Cheers,<br>> ><br>> > f<br>> ><br>><br>><br>><br>> --<br>><br>> Flávio Codeço Coelho<br>> registered Linux user # 386432
<br>> ---------------------------<br>> "Laws are like sausages. It's better not to see them being made."<br>> Otto von Bismark<br>> _______________________________________________<br>> IPython-user mailing list
<br>> <a href="mailto:IPython-user@scipy.org">IPython-user@scipy.org</a><br>> <a href="http://projects.scipy.org/mailman/listinfo/ipython-user">http://projects.scipy.org/mailman/listinfo/ipython-user</a><br>><br>
><br>><br></blockquote></div><br><br clear="all"><br>-- <br>Flávio Codeço Coelho<br>registered Linux user # 386432<br>---------------------------<br>"Laws are like sausages. It's better not to see them being made."
<br>Otto von Bismark