Comments on: Under the bonnet http://www.mysociety.org/2006/12/08/under-the-bonnet/ Relentless user-focus on civic websites Fri, 10 Feb 2012 01:00:32 +0000 hourly 1 http://wordpress.org/?v=3.0.5 By: Francis Irving http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1737 Francis Irving Mon, 02 Apr 2007 14:04:48 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1737 It would be a bit of work making it use a different database. Also, you would have trouble running the perl daemons it requires, if you can't install PostgreSQL. It would be a bit of work making it use a different database. Also, you would have trouble running the perl daemons it requires, if you can’t install PostgreSQL.

]]> By: Liam http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1733 Liam Mon, 26 Mar 2007 12:54:20 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1733 is it possible to make it work without postgress sql? that is all my server host offers. is it possible to make it work without postgress sql?
that is all my server host offers.

]]> By: Chris Lightfoot http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1610 Chris Lightfoot Mon, 11 Dec 2006 20:22:48 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1610 I benchmarked it (partly testing the database directly, partly with a full test of the whole signup process). We couldn't sustain the throughput we wanted with lots of database writers, but could with only one. It's not very surprising that the single-threaded implementation is more efficient. I benchmarked it (partly testing the database directly, partly with a full test of the whole signup process). We couldn’t sustain the throughput we wanted with lots of database writers, but could with only one. It’s not very surprising that the single-threaded implementation is more efficient.

]]> By: Dave Steinberg http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1609 Dave Steinberg Mon, 11 Dec 2006 18:50:55 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1609 Ah, nevermind commit_delay, I found this: http://archives.postgresql.org/pgsql-hackers/2002-10/msg00331.php Which suggests that gang-commit under load is self-tuning in PostgreSQL. I guess the question that remains is: is there evidence that supports the benefit of petsignupd? Ah, nevermind commit_delay, I found this:

http://archives.postgresql.org/pgsql-hackers/2002-10/msg00331.php

Which suggests that gang-commit under load is self-tuning in PostgreSQL. I guess the question that remains is: is there evidence that supports the benefit of petsignupd?

]]> By: Dave Steinberg http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1608 Dave Steinberg Mon, 11 Dec 2006 18:42:55 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1608 Did you consider adjusting commit_delay / commit_siblings in the database? Setting a delay of .25 (and a few siblings) seems like it would remove the need for petsignupd. Also, what sort of hardware is the database on? How are the disks organized? What sort of benchmarks did you run to test the setup? Did you consider adjusting commit_delay / commit_siblings in the database? Setting a delay of .25 (and a few siblings) seems like it would remove the need for petsignupd.

Also, what sort of hardware is the database on? How are the disks organized? What sort of benchmarks did you run to test the setup?

]]> By: Watching Them, Watching Us http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1606 Watching Them, Watching Us Sun, 10 Dec 2006 14:34:25 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1606 Ahah ! Mystery solved. I came to to the petition page following a Google search engine query which had picked up one of the names on the full list, but not in the most recent 500. Perhaps a "Display only the last 500" signatures link somewhere at the top of the form might be helpful ? Ahah ! Mystery solved. I came to to the petition page following a Google search engine query which had picked up one of the names on the full list, but not in the most recent 500.

Perhaps a “Display only the last 500″ signatures link somewhere at the top of the form might be helpful ?

]]> By: Chris Lightfoot http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1605 Chris Lightfoot Sun, 10 Dec 2006 13:04:15 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1605 Unless you click "show all signatories", the petitions site does only show the most recent 500 signatories. Unless you click “show all signatories”, the petitions site does only show the most recent 500 signatories.

]]> By: Watching Them, Watching Us http://www.mysociety.org/2006/12/08/under-the-bonnet/comment-page-1/#comment-1603 Watching Them, Watching Us Sun, 10 Dec 2006 11:29:37 +0000 http://www.mysociety.org/2006/12/08/under-the-bonnet/#comment-1603 Interesting, but the success of the currently most popular petition: "Scrap the planned vehicle tracking and road pricing policy (28207 signatures)" makes the front end web form virtually unusable - it is trying to display thousands of signatures at once and is therefore grinding my browser session to a halt. This is something which was fixed ok in PledgeBank by displayong only the last 500 signatures. Interesting, but the success of the currently most popular petition:

“Scrap the planned vehicle tracking and road pricing policy (28207 signatures)”

makes the front end web form virtually unusable – it is trying to display thousands of signatures at once and is therefore grinding my browser session to a halt.

This is something which was fixed ok in PledgeBank by displayong only the last 500 signatures.

]]>