PostgresSQL Halting System Boot

Matias Silva wrote:

Hi everyone, I'm totally new here and I got a little issue.
I just got done installing PostgreSQL 8.1. I have configured PostgreSQL
to start automatically upon system start via an init script. Problem
is that when CentOS boots it halts on "Starting PostgeSQL". I have
to press the <Enter> key to resume the boot. It doesn't time out either.

How do I fix this?

I check the PostgreSQL logs and the only thing that stands out is
"Do you want to choose a different one? [n]Your default context is
user_u:system_r:unconfined_t.". However when I search Google, I find
nothing.

This sounds like something to do with selinux, which Redhat build in to
their more recent offerings. I'm afraid I don't know enough about it to
give you specific advice, but you should be able to find out more by
checking the archives and googling.

--
Richard Huxton
Archonet Ltd

Matias Silva <matiassilva@gmail.com> writes:

How do I fix this?

I check the PostgreSQL logs and the only thing that stands out is
"Do you want to choose a different one? [n]Your default context is
user_u:system_r:unconfined_t.". However when I search Google, I find
nothing.

Dunno the actual cause. Someone else will.

Ok, so if you start the server again after booting you're going to get
prompted the same way. Try starting it with logging NOT directed to a
file.

$ su - postgres
$ pg_ctl start

Should prompt you. Now see if redirecting the null file to stdin lets
the server start without waiting.

$ pg_ctl start </dev/null

Or...

$ echo n | pg_ctl start

If any of these workarounds suffice, edit them into your
/etc/init.d/postgresql script.

HTH

--
-------------------------------------------------------------------------------
Jerry Sievers 305 854-3001 (home) WWW ECommerce Consultant
305 321-1144 (mobile http://www.JerrySievers.com/

Richard Huxton wrote:

Matias Silva wrote:

Hi everyone, I'm totally new here and I got a little issue.
I just got done installing PostgreSQL 8.1. I have configured PostgreSQL
to start automatically upon system start via an init script. Problem
is that when CentOS boots it halts on "Starting PostgeSQL". I have
to press the <Enter> key to resume the boot. It doesn't time out either.

How do I fix this?

I check the PostgreSQL logs and the only thing that stands out is
"Do you want to choose a different one? [n]Your default context is
user_u:system_r:unconfined_t.". However when I search Google, I find
nothing.

This sounds like something to do with selinux, which Redhat build in to
their more recent offerings. I'm afraid I don't know enough about it to
give you specific advice, but you should be able to find out more by
checking the archives and googling.

When logged in as root, I stop and start PostgreSQL by using the init script
and it doesn't halt on startup. So I decided switch to the postgres user
(su - postgres) and stat PostgreSQL (pg_ctl start -D ../data). Upon
start up it echos the log entry and waits for an enter key.

I notice the init script switches user.

start)
$ECHO_N "Starting PostgreSQL: "$ECHO_C
su - $PGUSER -c "$DAEMON -D '$PGDATA' &" >>$PGLOG 2>&1
echo "ok"
;;

Below is me switching user and starting the database and the blank
line is where I hit enter.

[root@CentOSVmware bin]# service postgresql stop
Stopping PostgreSQL: postmaster stopped
ok
[root@CentOSVmware bin]# su - postgres
-bash-3.00$ ./pg_ctl start -sD ../data
-bash-3.00$ LOG: database system was shut down at 2005-12-02 19:11:28 PST
LOG: checkpoint record is at 0/33B110
LOG: redo record is at 0/33B110; undo record is at 0/0; shutdown TRUE
LOG: next transaction ID: 588; next OID: 16385
LOG: next MultiXactId: 1; next MultiXactOffset: 0
LOG: database system is ready
LOG: transaction ID wrap limit is 2147484146, limited by database "postgres"

-bash-3.00$ ./pg_ctl stop -sD ../data
LOG: received smart shutdown request
LOG: shutting down
LOG: database system is shut down
postmaster stopped
-bash-3.00$ exit

Now that I know whats causing it how do I stop it, I don't know enough stop it.

Thanks for your time,
Matt

Matias Silva <matiassilva@gmail.com> writes:

I just got done installing PostgreSQL 8.1. I have configured PostgreSQL
to start automatically upon system start via an init script. Problem
is that when CentOS boots it halts on "Starting PostgeSQL". I have
to press the <Enter> key to resume the boot. It doesn't time out either.

I notice the init script switches user.

start)
$ECHO_N "Starting PostgreSQL: "$ECHO_C
su - $PGUSER -c "$DAEMON -D '$PGDATA' &" >>$PGLOG 2>&1
echo "ok"
;;

Hm, where did this init script come from? Last I heard, CentOS was a
direct ripoff^H^H^H^H^H^Hrepackaging of Red Hat, but the RH initscript
for Postgres doesn't look like that.

In any case, the Postgres community isn't in a position to support other
vendors' repackagings of Postgres. I suggest trying the initscript from
the PGDG RPMs (if not the entire RPM package...) --- that one we can at
least answer questions about.

regards, tom lane

On Sat, Dec 03, 2005 at 12:38:16PM -0500, Tom Lane wrote:

Matias Silva <matiassilva@gmail.com> writes:

I just got done installing PostgreSQL 8.1. I have configured PostgreSQL
to start automatically upon system start via an init script. Problem
is that when CentOS boots it halts on "Starting PostgeSQL". I have
to press the <Enter> key to resume the boot. It doesn't time out either.

I notice the init script switches user.

start)
$ECHO_N "Starting PostgreSQL: "$ECHO_C
su - $PGUSER -c "$DAEMON -D '$PGDATA' &" >>$PGLOG 2>&1
echo "ok"
;;

Hm, where did this init script come from? Last I heard, CentOS was a
direct ripoff^H^H^H^H^H^Hrepackaging of Red Hat, but the RH initscript
for Postgres doesn't look like that.

That is the init script distributed with postgresql -
contrib/start-scripts/linux.

That's the one I use in production and it works just fine.

In any case, the Postgres community isn't in a position to support other
vendors' repackagings of Postgres. I suggest trying the initscript from
the PGDG RPMs (if not the entire RPM package...) --- that one we can at
least answer questions about.

The original error smelled like an SELinux issue, rather than anything
to do with the startup scripts, but I'm not that familiar with SELinux
(I turn it off as part of the installation) so I can't say for sure.

Cheers,
Steve