Random internal Server Error when loading multiple images

I noticed a problem after migrating a contentbox Site to the latest version 5.2.0.
The site is basically used as a webshop so a lot of images might be included on each site.
When loading a site I noticed that not all images will load every time. After reloading the page the amount of images that will load is often times higher, however images that were loaded previously might not load again.
After some investigation I noticed that every time an image is not loaded a ConcurrentModificationException will be thown. (Stacktrace below).
The images are all located inside the “modules_app/contentbox-custom/_content” folder. However the _content folder itself is a symlink, linking to a folder outside the project.
This setup worker fine before, the problems started after our upgrade to Contentbox 5.2.0.
As I could not upload a .txt file the stacktrace is located in the attached pdf. Please let me know if another format is needed.
stacktrace.pdf (35.8 KB)

I noticed that I might have missedout on a few details of theinstallation:
The project is not hostet by commandbox but by a standalone lucee installation.
OS: Debian 10
Java: Open JDK 11.0.7
Apache Tomcat: 9.0.35
Lucee: 5.3.8.206
Contentbox: 5.2.0
I noticed that another project which still uses Contentbox 5.0.0-rc has the same problem, aout halfof the images on a given sit will not load. However it israndom which images are affected.
Deployment via Docker.
Base-image: tomcat:9-jdk11-corretto

@Christian we have tracked this down thanks to @bdw429s . It’s a lucee issue when creating internal Ids under load. We thought this was fixed in Lucee at this point, but it seems it’s a regression. @bdw429s can explain it better than I can.

However, for now, I am working on a patch for it. It’s actually a patch for WireBox. So it can workaround this issue.

I have bumped it to make sure I can get a release out this week.

1 Like

Quick update:
Yesterday I deployed the latest Lucee version which fixes the issue.

Edit: Looks like I was a little to optimistic here. The error still occures.

Hi @lmajano @bdw429s ,
Is there any update on this yet? Is there a workaround or a solution in prospect? Is there a ticket for this where I can follow the status on this?
BR
Christian

The fix is on WireBox. You can test with coldbox@be for now until we release the parch

1 Like