Re: [DOCS] Re: [HACKERS] Outline for PostgreSQL book
От | Vince Vielhaber |
---|---|
Тема | Re: [DOCS] Re: [HACKERS] Outline for PostgreSQL book |
Дата | |
Msg-id | Pine.BSF.4.05.9910130527030.7254-100000@paprika.michvhf.com обсуждение исходный текст |
Ответ на | Re: [DOCS] Re: [HACKERS] Outline for PostgreSQL book (Vince Vielhaber <vev@michvhf.com>) |
Список | pgsql-hackers |
On Tue, 12 Oct 1999, Vince Vielhaber wrote: > > On 13-Oct-99 Bruce Momjian wrote: > > Frankly, if I can get away with having whole sections that point to URL > > files, so much the better. If I can just point them to the INSTALL > > file, and be done with it, great. > > The installation can be quick and painless. I'll explain more tomorrow. Ok it's tomorrow. This will apply mainly to the actual installation process, but will also apply to the instructions for the book. 1) Add to configure: --with-postgres-user= and --with-postgres-group= with the default being postgres:postgres. 2) When generating the makefiles, use the -g and -o parameters to install. These, of course, will come from configure. At this point, anyone can build PostgreSQL, but only root and the postgres user can install it. Either way, the ownership will be correct since we added the -g and -o to install during the configuration process. Currently, the only difference between the way I do it and the above is that I override INSTALL in Makefile.custom (which I believe Tom Lockhart tipped me off to). I build as myself and use sudo to install (as many other admins also do). 3) gmake 4) gmake install (or sudo gmake install) Alternately, steps 3 and 4 can be done this way: 3a) gmake > gmake.out 2>&1 & tail -f gmake.out 4a) gmake install >> gmake.out 2>&1 & tail -f gmake.out But I don't usually bother. The final step may not be necessary if steps 1 and 2 are done right. I do it just to make sure the directories have the right permissions: 5) cd /usr/local (or whatever the directory is just below postgres') chown -R postgres:postgres pgsql (substitute asnecessary). Then do whatever initdb's and createdb's etc as necessary/desired. Even doing step 5 (and doing the custom makefile), on a 450 P-III running FreeBSD 3.2-RELEASE, I have the entire system installed and running in about 20-30 minutes. So how's that? There may be some minor differences between systems, I seem to recall HP having some with install(1), but I believe there was a workaround script that fixed it for most things called bsd.install or something like that. I'll verify HPs install on systems from 8-10.2 later this mourning - think I have a couple SGIs I can check too. Vince. -- ========================================================================== Vince Vielhaber -- KA8CSH email: vev@michvhf.com flame-mail: /dev/null # include <std/disclaimers.h> Have you seenhttp://www.pop4.net? Online Campground Directory http://www.camping-usa.com Online Giftshop Superstore http://www.cloudninegifts.com ==========================================================================
В списке pgsql-hackers по дате отправления: