Hacker News Reader

Comment by ellimilial | Hacker News Reader

It can’t run embedded. Having no live parts is a real paradigm shift for so many scenarios.

What does "can't run embedded" mean exactly?

I've run Postgres in a pure RAM only configuration, never touching the disk except to start.

If you mean "it can't be compiled in to another application as a library", then yes that's true - it's not an embedded database.

Can it be contained within the application, like SQLite?

Or has it to run as a separate process?

"can't run embedded" means you have to have a dedicated DB server process. For anything I would use a database for this is a non issue, but if you want a DB in your mobile app (for example) it might be.

So if SQLite is typically run embedded in the client application, does this mean it's not supported to have multiple clients operating on the same SQLite DB concurrently, or is there some support for this?

This is covered early on in the article.

It’s sad when most people discuss things based on a one line title of something they haven’t read. Many questions here were answered in that article. I recently asked someone on a different site “didn’t you read the article?”. The obvious answer I alluded to was “No, I don’t have time.” Yet they managed to post about 15 different comments asking more questions about the article (title). I really wish there was a way to have a public forum with proper discourse with the caveat that people must read the material before they comment on it.

The technical blog of the public broadcaster of Norway has implemented exactly this: https://nrkbeta.no/2017/08/10/with-a-quiz-to-comment-readers...

The quiz module is open-source (https://nrkbeta.no/2017/03/06/our-comment-quiz-module-is-now...) and available on github https://github.com/nrkbeta/nrkbetaquiz

Thanks for the links. :-)

It's mentioned briefly in the article, but no instructions are given on how to properly configure busy_timeout and WAL to scale above more than a few concurrent writes per second.

It uses filesystem locking mechanisms to synchronize concurrent access from multiple processes. One writer, multiple readers.

Kind of - your web application needs to partially handle that, and SQLite also depends on the OS filesystem to handle concurrent access.

But if the metric is "need in most cases," then that's fine. You don't need an embedded database in most cases.

The other way around. For most problems you don't need a DB. Only for the subset of those that do you need a dedicated process/container/machine.

We have learnt to use full-fledged RDBMSs as a default because they proved really flexible and powerful over the last 30+ years. But they do have limitations and cost.

Reply on news.ycombinator.com