626c2723d7
Fix a severe performance regression on the posts/index page introduced
by 6ca42947.
Short answer: scan_query dynamically allocated a regex inside an
inner loop that was called thousands of times per pageload.
Long answer:
* The post index page checks each post to see if they're tagged loli/shota,
* This triggers a call to Post#tag_array for every post.
* Post#tag_array called scan_query to split the tag string.
* scan_query loops over the tag string, checking if each tag matches the
regex /#{METATAGS.join("|")}:/.
* This regex uses string interpolation, which makes Ruby treat as a
dynamic value rather than a static value. Ruby doesn't know the
interpolation is static here. This causes the regex to be reallocated
on every iteration of the loop, or in other words, for every tag in
the tag string.
* This caused us to do thousands of regex allocations per pageload. On
average, a posts/index pageload contains 20 posts with ~35 tags per
post, or 7000+ total tags. Doing this many allocations killed performance.
The fix:
* Don't use scan_query for Post#tag_array. We don't have to fully parse
the tag_string here, we can use a simple split.
* Use the /o regex flag to tell Ruby to treat the regex as static and
only evaluate the interpolation once.
Installation
It is recommended that you install Danbooru on a Debian-based system since most of the required packages are available on APT. Danbooru has been successfully installed on Fedora, CentOS, FreeBSD, and OS X. The INSTALL.debian install script is straightforward and should be simple to adapt for other platforms.
For best performance, you will need at least 256MB of RAM for PostgreSQL and Rails. The memory requirement will grow as your database gets bigger.
On production Danbooru uses PostgreSQL 9.4, but any 9.x release should work.
Use your operating system's package management system whenever possible. This will simplify the process of installing init scripts, which will not always happen when compiling from source.
Troubleshooting
These instructions won't work for everyone. If your setup is not working, here are the steps I usually recommend to people:
-
Test the database. Make sure you can connect to it using psql. Make sure the tables exist. If this fails, you need to work on correctly installing PostgreSQL, importing the initial schema, and running the migrations.
-
Test the Rails database connection by using rails console. Run Post.count to make sure Rails can connect to the database. If this fails, you need to make sure your Danbooru configuration files are correct.
-
Test Nginx to make sure it's working correctly. You may need to debug your Nginx configuration file.
-
Check all log files.
Services
Danbooru employs numerous external services to delegate some functionality.
For development purposes, you can just run mocked version of these
services. They're available in scripts/mock_services and can be started
automatically using Foreman and the provided Procfile.
Amazon Web Services
In order to enable the following features, you will need an AWS SQS account:
- Pool versions
- Post versions
- IQDB
- Saved searches
- Related tags
Google APIs
The following features requires a Google API account:
- Bulk revert
- Post versions report
IQDB Service
IQDB integration is delegated to the IQDBS service.
Archive Service
In order to access versioned data for pools and posts you will need to install and configure the Archives service.
Reportbooru Service
The following features are delegated to the Reportbooru service:
- Related tags
- Missed searches report
- Popular searches report
- Favorite searches
- Upload trend graphs
Recommender Service
Post recommendations require the Recommender service.
Cropped Thumbnails
There's optional support for cropped thumbnails. This relies on installing
libvips-8.6 or higher and setting Danbooru.config.enable_image_cropping
to true.