On Tue, 2 Oct 2001, Thomas Lockhart wrote:
> Not sure about that. I started to refresh the doc build procedure but
> stopped when I couldn't figure out where the docs would go afterwards.
> I *think* I now know where that would be, but we are not to the point
> where I can actually write files there. And if
> developer.postgresql.org doesn't share disks with my home directory
> and the cvs repository, I'm looking at having to do a few more tweaks
> to move these things around from whatever the best "build machine"
> would be.
Its all on the same hard drive, on the same machine ...
> btw, in the machine info I've had before today,
> "developer.postgresql.org" didn't show up. So I'll still claim to be
> confused over what machines, aliases, and virtual machines are coming
> online, how they relate to each other, what disks they will see, etc
> etc. To quote an old advertisement down here, "inquiring minds want to
> know...". ;)
okay, let's try and do a summary here that is bound to miss something,
several monthss back, we moved to doing "virtual machine hosting", where
our clients were provided an isolated environment for their domains ... it
meant that we could provide very specialized services as a client required
(jakarta-tomcat for some, openacs for others, etc) without
conflicts/sharing between them ...
a 'domain machine' doesn't have any partitions per se, so if you do a 'df
.' from anywhre on cvs.postgresql.org, it will always be the same 'drive'
now,if you look at http://www.postgresql.org, it is just a "portal" now to
the mirrors, and hte mirrors have been tightened up so that any that is
out of date by 48hrs is automatically drop'd ... Vince and gang are
working on making it into even more of a portal site then what it is now,
with links to articles about PgSQL, news about it, etc ... work in
The 'physical sites' for PgSQL are moving to:
On the main server, where everything is pulled from,
http://www.ca.postgresql.org points to /usr/local/www/www/html ...
http://www.postgresql.org itself is on a different machine, as it doesn't
contain any "real" information, its all going to be dynamically generated
out of the database ...
http://archives.<cc>.postgresql.org is the mailing list archives which is
mhonarc based, and uses UDMSearch to search through them ... again,
everything in there is automagically generated, but is pulled from the
main server ...
http://fts.postgresql.org is still the OpenFTS search of the archives, so
there are two ways of searching the archives ...
http://gborg.postgresql.org, still being worked on, is in its own 'virtual
machine', since I didn't want to worry about overlaps with CVS ...
then there is techdocs and developer ...
archives was split from the main site, mainly to reduce teh size of teh
mirror, to make it optional instead of required ... www.<cc> is going to
be, as far as I understand it, the 'released stuff', while developer.<cc>
is going to be stuff like the TODO lists and views into the CVS repository
and whatnot ...
Alot of it is still a work in progress, and once we get the portal done,
we should have better 'directions' to give to ppl as to where they want to
go, but the 'central site' has just gotten too un-manageable from an
admin, as well as mirror, point of view ...
Anything that anyone is going to need to modify will always be on the
central server, which, once we get the new mirror stuff in place, will be
just plain 'postgresql.org', but for now is 'cvs.postgresql.org' ... ftp
directories are under ~ftp and web directories are all off of
/usr/local/www ... those will not change ...
does that help any, in the short term?
In response to
pgsql-docs by date
|Next:||From: terry||Date: 2001-10-02 16:36:19|
|Subject: Table view from another database|
|Previous:||From: Bruce Momjian||Date: 2001-10-02 16:14:10|
|Subject: Re: CVS changes|
pgsql-hackers by date
|Next:||From: Stephan Szabo||Date: 2001-10-02 16:41:23|
|Subject: Re: What about CREATE OR REPLACE FUNCTION?|
|Previous:||From: Lamar Owen||Date: 2001-10-02 16:35:45|
|Subject: Re: cvs problem|