Based on Debian GNU/Linux, PostgreSQL, Apache and Python I'm building
a web app. with PS/PDF documents as data. Their sizes will vary
between a few KB and many MB.
The docs will be printed via CUPS, scanned via SANE and fetched via
HTTP or FTP. Additionally their pages will be displayed inside the
client browsers in order to change their order, delete or insert
single pages. A future use case *might* be that the docs are
downloadable.
The question is wether I should store a doc as a file inside the file
system, as a postgres large object or as a text field inside a
postgres table.
The file approach is simple (KISS!) and employment of command line
tools like psselect would be very easy. OTOH data consistancy is an
important point, since the document maintainers will work in parallel.
Since scalability is also an issue: Is it possible to distribute one
single postgres table across multiple HD partitions?
I would appreciate to see your recommendations or further points I
should consider.
What about Performance and Backup/Restore issues?
TIA
Stony