I just deployed my Django (1.6) project with gunicorn and Nginx.
It seems to be working fine but I have one page were I'm getting an HTTP 500 error and I can't find any details about the error anywhere.
How do I get gunicorn to show me errors?
Here's all I currently see in the log file when I hit the page giving me the error:
>tail gunicorn.errors
2014-02-21 14:41:02 [22676] [INFO] Listening at: unix:/opt/djangoprojects/reports/bin/gunicorn.sock (22676)
2014-02-21 14:41:02 [22676] [INFO] Using worker: sync
2014-02-21 14:41:02 [22689] [INFO] Booting worker with pid: 22689
...
2014-02-21 19:41:10 [22691] [DEBUG] GET /reports/2/
Here's my bash script I use to start gunicorn:
>cat gunicorn_start
#!/bin/bash
NAME="reports" # Name of the application
DJANGODIR=/opt/djangoprojects/reports # Django project directory
SOCKFILE=/opt/djangoprojects/reports/bin/gunicorn.sock # we will communicte using this unix socket
USER=reportsuser # the user to run as
GROUP=webapps # the group to run as
NUM_WORKERS=4 # how many worker processes should Gunicorn spawn
DJANGO_SETTINGS_MODULE=reports.settings # which settings file should Django use
DJANGO_WSGI_MODULE=reports.wsgi # WSGI module name
#echo "Starting $NAME as `whoami`"
# Activate the virtual environment
cd $DJANGODIR
source pythonenv/bin/activate
export DJANGO_SETTINGS_MODULE=$DJANGO_SETTINGS_MODULE
export PYTHONPATH=$DJANGODIR:$PYTHONPATH
# Create the run directory if it doesn't exist
RUNDIR=$(dirname $SOCKFILE)
test -d $RUNDIR || mkdir -p $RUNDIR
# Start your Django Unicorn
# Programs meant to be run under supervisor should not daemonize themselves (do not use --daemon)
exec gunicorn ${DJANGO_WSGI_MODULE}:application \
--name $NAME \
--workers $NUM_WORKERS \
--user=$USER --group=$GROUP \
--log-level=debug \
--bind=unix:$SOCKFILE \
--error-logfile /opt/djangoprojects/reports/bin/gunicorn.errors \
--log-file /opt/djangoprojects/reports/bin/gunicorn.errors
More info:
I'm starting/stopping gunicorn with this init.d script I copied and modified using sudo service reports start|stop|restart
:
>cat /etc/init.d/reports
#!/bin/sh
### BEGIN INIT INFO
# Provides: django_gunicorn
# Required-Start: $local_fs $network $remote_fs
# Required-Stop: $local_fs $network $remote_fs
# Default-Start: 2 3 4 5
# Default-Stop: 0 1 6
# Short-Description: Starts django_unicorn reports at boot time.
# Description: Starts django_unicorn reports at boot time.
### END INIT INFO
name=`basename $0`
dir="/opt/djangoprojects/reports"
cmd="${dir}/bin/gunicorn_start"
pid_file="/var/run/$name.pid"
log_file="${dir}/bin/reports.log"
get_pid() {
cat "$pid_file"
}
is_running() {
[ -f "$pid_file" ] && ps `get_pid` > /dev/null 2>&1
}
case "$1" in
start)
if is_running; then
echo "Already running"
else
echo -n "Starting ${name}... "
cd "$dir"
#sudo -u "$user" $cmd &>> "$log_file"
$cmd &>> "$log_file" &
echo $! > "$pid_file"
if ! is_running; then
echo "Unable to start; see $log_file"
exit 1
else
echo "[STARTED]"
fi
fi
;;
stop)
if is_running; then
echo -n "Stopping ${name}... "
kill `get_pid`
for i in {1..10}
do
if ! is_running; then
break
fi
echo -n "."
sleep 1
done
echo
if is_running; then
echo "Not stopped; may still be shutting down or shutdown may have failed"
exit 1
else
echo "[STOPPED]"
if [ -f "$pid_file" ]; then
rm "$pid_file"
fi
fi
else
echo "Not running"
fi
;;
restart)
$0 stop
if is_running; then
echo "Unable to stop, will not attempt to start"
exit 1
fi
$0 start
;;
status)
if is_running; then
echo "[RUNNING]"
else
echo "[STOPPED]"
exit 1
fi
;;
*)
echo "Usage: $0 {start|stop|restart|status}"
exit 1
;;
esac
exit 0
From your comment I think this is a config problem in your django site, not a matter of gunicorn log, logs will not show more than django send to it.
Here is an example of how you can configure django setting to send log to your file (instead of send it to admins by email as default):
Read configuring logging (it provide a very well explanations of log settings options) and study the file django/utils/log.py to configure django loggin to appears more detailed on gunicorn logs.
Also check this answer and this which provide setting examples to send logs errors directly to a file. And consider to use Sentry to handle log errors, as is recomended by django guys.
Hope this helps.
1. sending errors to the console
These are the
loggers
that usemail_admins
by default (seedjango/utils/log.py
):you would need to change the handlers to go to the
console
so that it appears in your gunicorn log rather than send emails withmail_admins
. Please note that it's not as chatty as whenDEBUG=True
.2. sending errors via
mail_admins
Also based on configuring logging, explicitly create a handler that calls
mail_admins
; e.g. based ondjango/utils/log.py
:This requires you to set email related
settings
.3. other solutions
If you were not looking for solution #1, then your question is a duplicate of: How do you log server errors on django sites
The simplest solution is to configure the variable ADMINS with email addresses of people who should get error notifications. When DEBUG=False and a view raises an exception, Django will email these people with the full exception information.
settings.py
Maybe you need also EMAIL_HOST and EMAIL_PORT if the right SMTP server is not
localhost
on port25
. This simple solution is good enough for trial production operation, otherwise it can produce suddenly too much emails.Short answer:
With following logging configuration, your errors will start showing up in Gunicorn output(undaemonized) or runserver even when DEBUG is False. They anyways should be showing up when DEBUG is True.
If you want to see the Django errors in gunicorn error log, run gunicorn with --capture-output.
http://docs.gunicorn.org/en/stable/settings.html#capture-output
Long answer
There are two confusions involved when logging:
runserver
provide better log thangunicorn
settings.DEBUG=True
provide better log thansettings.DEBUG=False
Any log record you see with runserver can be seen with Gunicorn too as long as you have appropriate logging configuration.
Any log record you see with DEBUG=True can be seen while DEBUG=False too as long as you have appropriate logging configuration.
You can see default Django logging configuration at:
https://github.com/django/django/blob/1.10.8/django/utils/log.py#L18
It looks like: (I have stripped out parts which don't concern this answer)
What this says is:
Send
django
logger log record to handlersconsole
andmail_admins
.Handler
console
has a filterrequire_debug_true
on it. When settings.DEBUG is True, then handlerconsole
sends/prints the log on the Stream (because oflogging.StreamHandler
).When settings.DEBUG is False, then handler
console
ignores the log message sent to it by loggerdjango
.If you want logs to be printed with DEBUG=False too, then add a
handler
and make loggerdjango
use it.Handler would look like:
And use this handler with logger
django
:You can see the entire snippet in short answer.
With this, the logs will be printed on stream irrespective of if you are using runserver or gunicorn.
If you want the logs to be shown in gunicorn error log, then you need to run gunicorn with --capture-output.
This configuration worked for me. Add
--capture-output --enable-stdio-inheritance
with gunicorn command like below.With this setup, do enable logging in this way
This way you will get to see the errors in the App as well.