The main design goals of Ubooquity are simplicity and ease of use.
That's why Ubooquity is packaged in a single file and does not require any external dependencies (apart from the Java runtime environment) to run.

By the way, although Ubooquity acts as a web servers and uses an internal database, everything exists in a unique process.
There is no daemon, no external database instance, no external web server.

My guess is that pluging Ubooquity to your own database would not make you gain much anyway, a few megabytes of RAM perhaps. Not worth the trouble.
As for the Apache web server, as you guessed, it wouldn't bring anything at all.

It happens to everybody. :)

No, 24 comics per page is a low value indeed.

I was asking because Ubooquity has to read thumbnails files when displaying a page (as well as a few other files like themes element if you use a custom theme). But that's not important.

Ubooquity also scans your shared folders and all the files inside at startup, to detect any new, modified or removed file. It probably uses a lot of file handles at this time.

But don't focus on that.

If you want to get to the bottom of the problem, the first step is to characterize it.

The total number of file handles available on your system is shared by all applications.

Start by getting this max number and list consumption by application. There is a tutorial here.

The numbers you'll get will be a good starting point to understand what happens exactly.

NB: the web server is embedded because the goal of Ubooquity is to be self-sufficient, a single jar to copy and launch. And using an external server would not make sense does not need much in that domain. The embedded server is a single class: https://github.com/NanoHttpd/nanohttpd/blob/master/core/src/main/java/fi/iki/elonen/NanoHTTPD.java

Thanks, got it.

This file is of type "7z", not "zip". Ubooquity cannot open it.

I looked at my own 7z options; you have to specify the output format, otherwise 7zip will produce 7z files.

This is done using the "-t" switch (e.g. "-tzip" to produce zip files).

No it does not have one, but it should not require one (Ubooquity does not need to keep a lot of files open).

I'll have to do some tests to try to find a potential file handler leak.

Do you have some specific settings ? (like a very high number of comics/books per page)

It's on the top of the list for version 1.10 (not the next one, but the one after).

That's strange indeed. In theory Ubooquity should be able to read any zip file, regardless of the settings you used.

I have been using 7zip myself to convert cbr to cbz for years without any problem. I'll have to take a look at may settings.

When Ubooquity fails to open a file and tries to guess its format, it uses "magic numbers". Ubooquity not able to recognize a zip file would mean that the header of the file is not a standard one. Quite strange.

Could you provide me one of your faulty files so that I can analyze it ?

I'll also try to reproduce the problem on my side with your 7z settings.

"mas bum files" ?

You mean like creating some kind of named groups that would include several shared directories ?

If so, why not simply reorganize your folders ?