Strange behaviour on 499 http error. about nginx-upload-module HOT 8 CLOSED

1. What version of nginx are you seeing this 499 error with?

2. Is it the base nginx-upload-module plugin from repo that you are using or a modified version?

from nginx-upload-module.

1. nginx -V

   nginx version: nginx/1.1.15
   built by gcc 4.1.2 20080704 (Red Hat 4.1.2-52)
   TLS SNI support disabled
   configure arguments: --prefix=/etc/nginx/ --sbin-path=/usr/sbin/nginx --conf-path=/etc/nginx/nginx.conf --error-log-path=/var/log/nginx/error.log --http-log-path=/var/log/nginx/access.log --pid-path=/var/run/nginx.pid --lock-path=/var/run/nginx.lock --http-client-body-temp-path=/var/cache/nginx/client_temp --http-proxy-temp-path=/var/cache/nginx/proxy_temp --http-fastcgi-temp-path=/var/cache/nginx/fastcgi_temp --http-uwsgi-temp-path=/var/cache/nginx/uwsgi_temp --http-scgi-temp-path=/var/cache/nginx/scgi_temp --user=nginx --group=nginx --with-http_ssl_module --with-http_realip_module --with-http_addition_module --with-http_sub_module --with-http_dav_module --with-http_gzip_static_module --with-http_stub_status_module --with-ipv6 --with-debug --add-module=../nginx_upload_module-2.2.1 --add-module=../nginx_eval_module-1.0.3 --add-module=../mod_zip-1.1.6
   

2. I use original upload module compiled from source, without any modifications.

Actually, 499 error is not the main point. The state file inconsistency is much more bothers me.

from nginx-upload-module.

Someone approached me with similar problem before. I need more information in order to figure out what happens. Please provide some relevant debug logs.

from nginx-upload-module.

I've found the root cause of the problem.
Client sends last chunk, sets timeout and waits for the response.
Upload module handles upload, removes state-file (this is important!) and proxies request to backend specified in upload_pass directive. If backend is overloaded and doesn't response during client's timeout client close connection and repeat the same request. But this request is treated as new upload because state-file already removed.
I think the possible solution is do not remove state-file right after upload processing but remove it later, after response from backend. What do you think about this?

from nginx-upload-module.

Yes, we can choose not to remove the state files at all. In that case you can remove them later by running find ... | xargs rm script from cron.

In order to implement that, just outcomment lines 1520-1524 in this file:
https://github.com/vkholodkov/nginx-upload-module/blob/2.2/ngx_http_upload_module.c

However, this will produce duplicate requests on backend side and you have to deal with them somehow (with proper implementation it shouldn't be difficult).

from nginx-upload-module.

Is it safe to delete state-file in ngx_upload_cleanup_handler? This will require to add another flag in ngx_upload_cleanup_t like finished.
What do you think about this?

from nginx-upload-module.

Depends on the client. If client doesn't do simultaneous request, this should be okay. Otherwise it could lead to similar problems.

from nginx-upload-module.

I propose to add flag (finished, for example) indicating that file fully uploaded and to delete state-file in ngx_upload_cleanup_handler in case of finished && !aborted.
I do not understand why simultaneous requests will lead to the problems, could you please explain a bit more detailed?

from nginx-upload-module.