Sphinx or Yandex.Server?

Question

MySQL

Actually a subject. There is a lot of text (4000 thousand articles, 8 thousand titles).
Everything is spinning using the mysql DBMS on linux 2.6.
Interested in how both products deal with memory consumption and output quality.

Answer 1

4000 or 4.000.000 articles?
You can set either one or the other.
Yandex server is more capricious in settings and heavier with documentation. A year ago, there were problems with encoding when indexing through mysql, somehow we managed to get around it.
But there are also some bonuses, in particular, ready-made issue with highlighting. For a small number of documents, Yandex is probably convenient. The main difference is that Yandex when searching saves you from additional. access to the database of documents.
The Sphinx is much easier to climb, there are no problems with the documentation (there is a live forum) and more flexible in the settings.
For really large databases, the sphinx will give Yandex a hundred points ahead due to the ability to store most of the database on the screw (Yandex had problems).
the lack of a sphinx in the need to climb into the sql database with handles when searching (in the last beta, it became possible to store a selection in the sphinx database, but for that it is a beta to be inconvenient. on small databases it is usable, on large it slows down).
From living examples: on the basis of orders of hundreds of millions of records, a sphinx was chosen. Yandex stupidly could not survive on one machine (it did not find a mode that allows it to minimize the amount of memory).

Answer 2

Sphinx, Yandex.Server is very picky about the OS and settings (it takes a lot of time to file it)

Answer 3

I don’t remember if Yandex.server has the ability to index the database directly, unlike the Sphinx. Emnip it only http crawl and another file on the screw can index. for this, decide that it is easier for you to set the sphinx on the base or Yandex.server on the site.
in general, proceed from the fact that you actually want to index - site pages or a column in a database table, and then screw in what suits you.
something like this. I installed Yandex.server a couple of times, I don’t remember any difficulties. Sphinx did not set, I can not compare.

Answer 4

In Ya.S, you can index the MySQL database and any other database via ODBC.
I raised both the Sphinx and Y.S., of course, the relevance of Y.S. is higher, but it takes a very long time to index a site of 450,000 pages, the approximate time is 2 days, and during indexing it needs 2 gigabytes on the screw to be free (for my amount pages). And when he starts to merge indexes, the server hangs up, the hoster constantly complained that my server was eating more RAM than it could (2Giga). In short, it's up to you, but with 4M articles, if I put Y.S, then on a separate server.