yes, i think once the RAM was consumed the server ground to a halt where it appeared to crash. so, i believe it was a soft crash instead of a hard crash. i used "top" to monitor the memory and processes and the soft crash occurred when the RAM was exhausted. so i created a swap file to extend the RAM using this series of CLI/sh commands:
df -h free -m swapon --show fallocate -l 2G /swapfile chmod 600 /swapfile mkswap /swapfile swapon /swapfile #under /etc/fstab, add the line: /swapfile swap swap defaults 0 0 swapon --show free -m df -h which did the trick as i still watched under top. therefore, creating the swap file did the trick. i also updated my uwsgi initialization file to allow for more processes, or: [uwsgi] chdir = /opt/web2py_apps/web2py module = wsgihandler:application master = true processes = 5 threads = 2 thread-stacksize = 512 uid = web2py gid = web2py socket = /run/uwsgi/web2py.sock chown-socket = web2py:nginx chmod-socket = 660 vacuum = true thank you for your help. lucas On Tuesday, October 17, 2023 at 3:43:31 AM UTC-4 Tom Clerckx wrote: > Maybe you should review the 'high calculation/high plotly' pages for > memory usage? > As the 'OOM killer' kicked in, it looks like the process is using too much > memory. > > One thing to verify this is to monitor the VmRSS of the processes involved > (VmRSS is one of the fields in the output when you do 'cat > /proc/<pid>/status') > VmRSS gives you an idea about the amount of RAM memory your process is > using. > > If it's growing out of proportion, you're claiming too much memory and the > process will be killed by the OOM-killer. > > > On Monday, October 16, 2023 at 8:15:20 PM UTC+2 lucas wrote: > >> hello one and all, >> >> i'm using web2py 2.25.1 with uwsgi 2.0.22 which was downloaded and >> compiled using pip3 under python 3.10.12 under ubuntu 22.04. the nginx is >> apt installed version 1.18. >> >> with that being said, if i install and run my high calculation / high >> plotly plot pages about 6 to 12 times, the whole damn system comes to a >> crash. hard crash. this also seems to be the case under redhat / amazon >> linux 2023 similar setup and config. >> >> so i tried crashing the system under just "python3 web2py.py -i 0.0.0.0 >> -p 80 --no_gui" and also under "uwsgi --http :80 --chdir >> /opt/web2py_apps/web2py -w wsgihandler:application" but i could NOT get >> either to crash. it is only when i nginx daemon it with uwsgi and web2py >> layered/socketed under nginx that it DOES crash hard. i think i found >> under the systemctl logs for uwsgi a logging of the crash, or: >> >> Oct 16 13:48:30 iquanta.info systemd[1]: uwsgi.service: A process of >> this unit has been killed by the OOM killer. >> Oct 16 13:48:30 iquanta.info uwsgi[1264]: [emperor] *** RAGNAROK EVOKED >> *** >> Oct 16 13:48:30 iquanta.info uwsgi[1265]: SIGINT/SIGTERM >> received...killing workers... >> Oct 16 13:48:30 iquanta.info uwsgi[1264]: Mon Oct 16 13:48:30 2023 - >> [emperor] stop the uwsgi instance web2py.ini >> Oct 16 13:48:30 iquanta.info uwsgi[1265]: Mon Oct 16 13:48:30 2023 - >> received message 0 from emperor >> Oct 16 13:48:31 iquanta.info uwsgi[1265]: worker 5 buried after 1 seconds >> Oct 16 13:48:32 iquanta.info uwsgi[1265]: worker 1 buried after 2 seconds >> Oct 16 13:48:32 iquanta.info uwsgi[1265]: worker 2 buried after 2 seconds >> Oct 16 13:48:32 iquanta.info uwsgi[1265]: worker 3 buried after 2 seconds >> Oct 16 13:48:32 iquanta.info uwsgi[1265]: worker 4 buried after 2 seconds >> Oct 16 13:48:32 iquanta.info uwsgi[1265]: goodbye to uWSGI. >> Oct 16 13:48:32 iquanta.info uwsgi[1265]: VACUUM: unix socket >> /run/uwsgi/web2py.sock removed. >> Oct 16 13:48:33 iquanta.info uwsgi[1264]: Mon Oct 16 13:48:33 2023 - >> [emperor] removed uwsgi instance web2py.ini >> Oct 16 13:48:34 iquanta.info uwsgi[1264]: Mon Oct 16 13:48:34 2023 - The >> Emperor is buried. >> Oct 16 13:48:34 iquanta.info systemd[1]: uwsgi.service: Failed with >> result 'oom-kill'. >> Oct 16 13:48:34 iquanta.info systemd[1]: uwsgi.service: Consumed 10.241s >> CPU time. >> Oct 16 13:48:34 iquanta.info systemd[1]: uwsgi.service: Scheduled >> restart job, restart counter is at 1. >> Oct 16 13:48:34 iquanta.info systemd[1]: Stopped uWSGI Emperor service. >> Oct 16 13:48:34 iquanta.info systemd[1]: uwsgi.service: Consumed 10.241s >> CPU time. >> >> with not much to say under the nginx logs. so, does anyone know what i >> can be dealing with? thank you in advance, lucas >> > -- Resources: - http://web2py.com - http://web2py.com/book (Documentation) - http://github.com/web2py/web2py (Source code) - https://code.google.com/p/web2py/issues/list (Report Issues) --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to web2py+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/web2py/9513d032-6e23-4f53-b29f-8438322c0725n%40googlegroups.com.
