PostgreSQL pain points [LWN.net]

PostgreSQL pain points

Posted Apr 8, 2014 20:14 UTC (Tue) by walex (guest, #69836)
Parent article: PostgreSQL pain points

The funny thing about this report is that the Postgres researcher had made almost the same type of complaints about UNIX and Ingres in 1981, 33 years ago:

%A M. R. Stonebraker
%T Operating system support for database management
%J CACM
%V 24
%D JUL 1981
%P 412-418
%K data base

Another funny detail is that adaptive advising about access patterns (adaptive read head, adaptive write behind) were part of the original UNIX design on the PDP-11 well around 35-40 years ago.

Also good guesses for semi-automatic advising could work in most cases for most programs by looking at the flags/modes with which a file is opened in fopen or even open as well as the pattern of IO operations on the file like UNIX 25-40 years ago.

to post comments

PostgreSQL pain points

Posted Apr 9, 2014 21:04 UTC (Wed) by bcopeland (subscriber, #51750) [Link]

>Also good guesses for semi-automatic advising could work in most cases for most programs by looking at the flags/modes with which a file is opened in fopen or even open as well as the pattern of IO operations on the file like UNIX 25-40 years ago.

There was a paper a while ago where some researchers looked at the filenames for such hints and got quite good results (e.g. an .mpg is mostly read-only and streamed, a .sqlite file may have seeky writes, and so on.) Cute hack.