On Tue, Dec 4, 2012 at 3:38 PM, Gunnlaugur Þór Briem <[email protected]> wrote: > On Tuesday, 4. December 2012 at 21:32, Gunnlaugur Thor Briem wrote: > > I happened to copy-paste a run where exactly the first four worked :) --- > don't let that confuse you. The selection is random, though it seems biased > in favor of the lower byte counts. Here's another sample:
[...] > If the pipe were working correctly, we ought to see 24 messages, with all > byte counts 1030 up to 1041 and then 1041 back down to 1030. Most of them > seem to get snatched by the evil pipe dragons. possibly related/similar: http://bugs.python.org/issue11907 ...not Mac OSX (FreeBSD), and not pipe buffer filled (socket buffer), but interestingly, the OP was using uWSGI, trying to solve the same problem, and 2048-ish buffers were mentioned... per the thread, try: logging.raiseExceptions = False "[...] can't reproduce this on Linux even when sending messages of > 16384 on the Unix socket" "[...] the error message is being printed by the handleError() method of the handler" "[...] the exception is indeed not propagated to the application [...] I was confused by seeing the traceback in my uWSGI log file" "[...] set logging.raiseExceptions to False" "[...] why I started using syslog [...] log into a single file from multiple processes, but it seems to be showing up as too much trouble." -- C Anthony _______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
