bulk_add_csv 504 gateway timeout after exactly 60s

[Kodtiz3D]

Server Software

• Shimmie version: 2.12.2-20260225-92e0fe6
• Install method: Docker
• Server OS: Linux - TrueNAS SCALE
• Database: sqlite 3.46.1
• Web server: nginx 1.26.3
• Theme: lite
• Nice URLs: enabled

Client Software

• Device: Desktop Linux
• Browser: Firefox 148.0-1 and Chromium 145.0.7632.109

What steps trigger this bug
Get a sufficiently large .csv file. My testing showed 70+ rows was enough. Might depend on how powerful the host is though.
Try to import through the web interface and wait 60 seconds.
After 60 seconds I get a 504 Gateway timeout.

What did you expect to happen?
The import to still happen in the background.

What actually happened?
No files were added/modified.

I can't seem to find any configuration files to change the timeout, and looking at the main.php file here it's supposed to disable the timeout during this process. At least that's what I expect to happen. In the mean time, is there a way to run/invoke the script in the background as a workaround? Hopefully directly on the host, rather than through the web interface.
With a smaller csv of about 30-60 rows long I was able to get it working as intended. And a much larger csv just breaks it outright. With a HTTP 500 error.

[Kodtiz3D]

I did try to run php -f /app/ext/bulk_add_csv/main.php from within the Docker container, and it complained that Class "Shimmie2\Extension" not found in /app/ext/bulk_add_csv/main.php:14
So it would need more work than just that. I found the php socket timeout setting, but that didn't help. It must be nginx (or something else I'm not familiar of) doing the timeout.

[sherman2444]

I tried modifying /etc/nginx/nginx.conf in the container to fix this, but it's recreated at container start by /app/.docker/run.php. I ended up copying run.php out of the container, modifying it and building a new image from Dockerfile:

FROM shish2k/shimmie2:2

COPY ./run.php /app/.docker/run.php

# Set PHP max_execution_time:
RUN sed -i 's/max_execution_time = 30/max_execution_time = 600/g' /etc/php/8.4/fpm/php.ini

I added 3 options to the config string that run.php writes to /etc/nginx/nginx.conf under location ~ \.php$:

    fastcgi_read_timeout 600s;
    fastcgi_send_timeout 600s;
    fastcgi_connect_timeout 600s;

I no longer get 504 errors when running bulk upload. Hope this helps!

[Kodtiz3D]

Good news! That did work to get past the timeout. Bad news. Now I get a 500 error when loading a 177 line csv about 124kB in size.
Looking into the logs I have, it seems to be getting an OoM error with PHP.
[error] 48#48: *17 FastCGI sent in stderr: "PHP message: PHP Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 7340161 bytes) in /app/core/Media/MimeType.php on line 172" while reading response header from upstream

While I could increase the allocated memory that PHP has access to, that will only be a bandaid solution. As I have a much larger csv I wish to push to my server. About 11.6MB. I guess I'll have to learn proper SQLite DB manipulation to get what I want here.