Hello guys. I'm trying to run a python app in emperor mode and i get the error of the subject. In the log, the backtrace is empty, there is no more lines to add. If ran in the console with a very similar ini file and works like a charm.
My vassal ini is:
[uwsgi]
dir = /var/www/KRAKEN/public/website
chdir = %(dir)
master = true
processes = 1
socket = socket.sock
chmod-socket = 666
enable-threads = true
vacuum = true
plugins = python
virtualenv = %(dir)
binary-path = %(dir)/bin/uwsgi
mount = /website=app.py
manage-script-name = true
logto = %(dir)/uwsg.log
The one that runs fine:
[uwsgi]
dir = /var/www/KRAKEN/public/website
chdir = %(dir)
master = true
processes = 1
socket = socket.sock
chmod-socket = 666
plugin = python
enable-threads = true
vacuum = true
plugins = python
virtualenv = %(dir)
mount = /website=app.py
manage-script-name = true
emperor.ini:
[uwsgi]
emperor = /etc/uwsgi/vassals
uid = gabriel
gid = http
UWSGI version 2.0.12 running in Archlinux for ARMv7.
Tips?
Plese provide a backtrace otherwise this is quite hard to debug. Also please post the diff between vassal.ini and the one that works correctly.
@xrmx Hi Ricardo. Thanks.
I only get:
Mar 18 11:44:09 KRAKEN uwsgi[350]: !!! uWSGI process 5447 got Segmentation Fault !!!
Mar 18 11:44:09 KRAKEN uwsgi[350]: *** backtrace of 5447 ***
Mar 18 11:44:09 KRAKEN uwsgi[350]: *** end of backtrace ***
Regarding the script, i believe it has little to do. If i run in the webapp directory (uwsgi --ini file.ini), it runs fine, if i move it to /etc/uwsgi/vassals and run it with the emperor systemd service, i get the segmentation fault.
@xrmx
This is the content of /var/lib/systemd/coredump
[New LWP 7639]
Core was generated by `/usr/bin/uwsgi --ini website.ini'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x0000123c in ?? ()
This happens when running as SU and also when trying to execute as a service.
@xrmx Here is the strace of the command
systemctl start emperor.uwsgi
that is the systemd service from Archlinux for starting the emperor mode:
@GAZ082 again, without a backtrace or at least a reproducer there's no way one can understand what is going on there.
@xrmx I tried my best with _strace_ and the backtrace shown by uWSGI is what already posted.
How can i provide you with a better backtrace?
Thanks!
@xrmx the only way to get this working is to login to the server console and run the command manually. If i do a crontab with the command to run every reboot, it does not work, same error. I hate not being able to give you more info, this really pisses me off.
Most helpful comment
@xrmx the only way to get this working is to login to the server console and run the command manually. If i do a crontab with the command to run every reboot, it does not work, same error. I hate not being able to give you more info, this really pisses me off.