_____/ On Thu 26 Jan 2006 02:09:36 GMT, [Chetan Kunte] wrote : \_____
Hi all -
My blog with less than a 1000 posts weighs slightly over 0.5MB in a
gzip format and is over 3MB (bare minimum, without any extra tables
like stats and other optional stuff). I am worried that sooner or
later the database is going to be a big fat pig that will refuse to
fly in one piece (backup and restore). On a lowest common denominator
level, this could get a lot worse for a lot of users: dial-up, hosts
that do not offer a command line, etc. As a user myself (with a decent
DSL connection and host) still dread the idea of squashing the db and
uploading afresh from the backup for the fear of not completing the
You have raised a good point, particularly when it comes to slow connections
and restrictive Web hosts. Nevertheless, many hosts continue to permit the
cron job functionalities while denying command-line access. My host is
among them, so one can set up a cron job to sqldump the database onto the
Web server automatically.
If you fear the idea of restoring a database, then your fear is in vain. We
are all victims of paranoia at times. There were only two situations I can
recall where database restoration went awry or was lossy. With WordPress, a
restoration I once attempted led to an issue with a table that's associated
with PDF creation settings. Settings were easy to restore manually though.
With phpWiki, I once uploaded a database with excess verbage. Everything
else was fine.
Remember that a failed upload can be superseded by a second attempt. If
needed, you can confer somebody else for help, provided you keep the raw
Are you guys thinking of something that could ease this pain and keep
the pig always in air? :)
If you habitually save your database off-server, then you could just store
the 'diff' between the current database state and yesterday's state. For
example, consider the following cron job:
0 0 * * 0,2,4,6 mysqldump --user <username> --password=<your_pass>
<your_database> > ~/tmp/blog.dump;
0 0 * * 1,3,5 mysqldump --user <username> --password=<your_pass>
<your_database> > ~/tmp/blog2.dump;
5 0 * * * diff ~/tmp/blog.dump ~/tmp/blog2.dump >>
This does not obviate the need for full backups, but it keeps you on the
'safe side' more frequently.
In a layman's terms, I was thinking of an idea: how about if the
database can be backed-up and restored incrementally like "per year"
or "per number of posts" in addition to the default option of a full
backup and restore. I seriously think the "per year" or "per number of
posts" concept is a good idea (provided of course the core contents of
a database are not modified in between versions. Do you think I'm
I think that Scott (skippy) might wish to consider this. He's the WP 'backup
man'. Light backups ought to help scalability, yet without potential clutter
(maybe "advanced" tab).
As always, a huge WordPress fan.