Thumbnails over I2P produce 500 errors

YaCy
1

Hello,

I’ve been using YaCy to crawl I2P, and it works very well overall. However, I’ve run into an issue: the image search doesn’t seem to work at all—no thumbnails are loading.

When I check the browser console, all image URLs return a 500 server error. Strangely, when I visit those URLs directly in the browser, the images load without any problem. For example:

http://leftychan.i2p/music/src/1615683388558.png

This loads fine when accessed directly, but fails through YaCy’s image search.

I should note that, of course, I have significantly increased timeouts due to crawling over I2P, so it’s not a simple timeout issue.

image
image603×141 12.3 KB

2

Welcome @janio55

Do you run yacy and i2p on Windows?

3

Thank you for your response, No I run Yacy and I2P both inside of docker.

4

Hi. I don’t really know much about docker.

Do you know how to go from a github repository to Docker image?

Try this code!

https://github.com/smokingwheels/yacy_search_server

I had a go at docker it but it failed.

5

I might be able to but what difference is their that might solve this issue?

6

Three source files were patched (ViewImage, HTTP client handling, proxy settings) to improve robustness behind proxies.

Files touched (summary)

  • htroot/ViewImage.java
    Improves image retrieval and decoding reliability by adding retry logic and tolerating slow or partial streams, which are common when images are fetched through proxies or tunnels.

  • HTTPclient–related class (e.g. HttpClient usage / wrapper)
    Adjusts request handling to behave more defensively when connections are mediated by a reverse proxy (timeouts, retries, and stream availability assumptions).

  • Proxy configuration / settings handling
    Minor changes to how proxy settings are respected during outbound HTTP requests, helping ensure consistency between normal page fetches and auxiliary resources (images, icons, peer checks).

This is not a formal upgrade and has no guaranteed upgrade path.

You would have to export and import your content to the new install.

7

So i got it fixed with the docker setup I am using.
I hat to enable a few things. first one is inside the Web Structure Index i enabled use citation reference index (lightweight and fast).
As well as enable process_sxt inside the Solr Schema Editor

I dont know why that isnt enabled by default