Compiling/Installing as a non-admin user

On 11/1/06, Ritesh Nadhani <riteshn@gmail.com> wrote:

Hello All

Me and my professor are planning to work upon machine learning in
postgresql over tsearch2. So I have some questions:

We have a server where Postgresql is running without any problem with
postgres username and admin rights. I have a user account in that
server. I plan to compile and run another postgresql for our testing so
I was thinking of how to do that? My prior knowledge of using postgresql
has always been as admin where I have full rights.

As I see, using the default MAKE for postgresql will set the data
directory etc. in /usr/local/data etc which I dont have access to as a
user.

So I would like to compile and run postgresql as a normal user with
every thing like data kept in my usr directory. I should be able to run
the instance over separate port and can start and stop it.

Basically, I want to run the server as in user mode....

How should I configure the MAKE and INSTALL in this circumstances? What
are your suggestions....

Ritesh

---------------------------(end of broadcast)---------------------------
TIP 3: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faq

Gurjeet Singh wrote:

You are talking about /usr/local/data; so I assume that you are trying
this
on Linux or some other Nixen.

You just need to use the --prefix option to configure... Here's what I
typically do:

Download/'CVS checkout' the sources. Enter the source directory, and
invoke
configure like this:

./configure --enable-debug --prefix `pwd`/db CFLAGS=-O0

And the run 'make' and 'make install'. This will install the data in your
<sources_dir>/db/data.

And then of course you'll need to remember to set your port to something
other than 5432, and tweak your PATH, PGPORT etc or define some
aliases/wrappers so you don't end up running against the default
installation. Oh, and you'll need to tweak the startup scripts and
logging configuration so you get logs somewhere useful.

I think Tom Lane has a script that lets him switch between different
installations (versions in his case). I only tend to have two versions
active at any one time, so I just define an alias for psql.
alias psql82='/usr/local/pgsql82/bin/psql -p 5434'
On the rare occasion when I run an 8.2 createdb I need to remember to
put the port number in manually of course.

HTH

Gurjeet Singh wrote:

You are talking about /usr/local/data; so I assume that you are trying
this
on Linux or some other Nixen.

You just need to use the --prefix option to configure... Here's what I
typically do:

Download/'CVS checkout' the sources. Enter the source directory, and
invoke
configure like this:

./configure --enable-debug --prefix `pwd`/db CFLAGS=-O0

And the run 'make' and 'make install'. This will install the data in your
<sources_dir>/db/data.

And then of course you'll need to remember to set your port to something
other than 5432, and tweak your PATH, PGPORT etc or define some
aliases/wrappers so you don't end up running against the default
installation. Oh, and you'll need to tweak the startup scripts and
logging configuration so you get logs somewhere useful.

I think Tom Lane has a script that lets him switch between different
installations (versions in his case). I only tend to have two versions
active at any one time, so I just define an alias for psql.
alias psql82='/usr/local/pgsql82/bin/psql -p 5434'
On the rare occasion when I run an 8.2 createdb I need to remember to
put the port number in manually of course.

HTH

On 11/1/06, Ritesh Nadhani <riteshn@gmail.com> wrote:

Hello All

Thanks for all the suggestions. I was able to configire and compile it.
I set the data directory to one of my directories, use 'initdb'
successfully. I am also able to start the pgsql and can see that it is
running on port 4444 as I had configured.

Now the problem is when I try to create a db using createdb, I get the
error:

createdb: could not connect to database postgres: FATAL: role "ritesh"
does not exist

I am not sure what is that?

Doing a google gave me:
http://pgfoundry.org/pipermail/pgcluster-general/2006-May/000699.html

You must be postgres user to run this. Another option might be

passing > a user switch .

I cannot run with postgres as I dont have access and I am not sure how
to use the user switch? Any help?

I guess once I am through with this, I am successful!

Ritesh

Richard Huxton wrote:

Gurjeet Singh wrote:

You are talking about /usr/local/data; so I assume that you are trying
this
on Linux or some other Nixen.

You just need to use the --prefix option to configure... Here's what I
typically do:

Download/'CVS checkout' the sources. Enter the source directory, and
invoke
configure like this:

./configure --enable-debug --prefix `pwd`/db CFLAGS=-O0

And the run 'make' and 'make install'. This will install the data in

your

<sources_dir>/db/data.

And then of course you'll need to remember to set your port to something
other than 5432, and tweak your PATH, PGPORT etc or define some
aliases/wrappers so you don't end up running against the default
installation. Oh, and you'll need to tweak the startup scripts and
logging configuration so you get logs somewhere useful.

I think Tom Lane has a script that lets him switch between different
installations (versions in his case). I only tend to have two versions
active at any one time, so I just define an alias for psql.
alias psql82='/usr/local/pgsql82/bin/psql -p 5434'
On the rare occasion when I run an 8.2 createdb I need to remember to
put the port number in manually of course.

HTH

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly