I have found the source of the issue and it seems to be a potential bug so I'm going to open an issue on GitLab.
When uwsgi is initially started via `/etc/init.d/supervisor restart`, everything is fine, and I can use the above described method of creating tags. The problem seems to surface after calling various API endpoints (probably other web endpoints also) that cause a particular 500 error. One of these endpoints that causes a 500 is `/document/<document_pk/metadata/` (GET). Once a call to the above endpoint is made the following error is logged mutliple times (usually 5 or 6 times) in `/var/log/uwsgi/uwsgi.log`: navigation.classes <16376> [WARNING] "resolve() No request variable, aborting menu resolution" After this call is completed, *the server is left in a broken state and will keep returning 500s only for particular routes until uwsgi is restarted.* Not all routes are affected by this though, so many routes/endpoints will continue working. It just takes so trial and error to figure out which endpoints fail once in a broken state. This error also surfaced in this issue <https://gitlab.com/mayan-edms/mayan-edms/issues/244> on gitlab a few month ago. However, in that case it seemed to be caused by virtualbox issues unrelated to code. The machines I'm running mayan on all have at least 2gb of memory (most are 4gb+), have multiple cores, are dedicated to running Mayan, and are not running any virtualization software. So it feels safe to rule out the above cause. My 'workaround' for now is to just not use the endpoints that cause this problem. If anyone has any info on this or has experienced the same errors, please let me know, any info is useful. I also hope that this post is useful to someone out there. :) -- --- You received this message because you are subscribed to the Google Groups "Mayan EDMS" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
