Is MongoDB + Sphinx + 100Gb worth it?

Question

Sphinx

Good afternoon.
Until sphinx 3.0 is released, the only way to connect sphinx and mongo is with xmlpipe The
question is, how reliable and performant is this solution on 100Gb+ of data?

The database stores information consisting of ID + text.
+ texts can have variable characteristics... links, pictures (url), various notes.
In principle, you can presume in advance what attributes are generally possible for texts and push it all into the format of one or related tables ....
Perhaps you should abandon mongo in favor of pg or mysql-hs ?

Answer 1

taking into account the fact that the texts do not change, just make the main index and the delta index, it will be very productive.

Answer 2

Why the only one? I am using sphinxql.

Answer 3

Perhaps, by the way, you don’t need to import mongo-> sphinx. And make RT-index and duplicate requests. When filling in the data, fill in parallel RT and mongo.... The question here will be only in the regeneration of the index when it is damaged....
Or does such a solution have its drawbacks?