postgresql wont start!

[tranquilcool]

for some reasons postgresql would not start.

Code: Select all

/etc/init.d/postgresql.10 start

/run/postgresql: correcting mode
 * Starting PostgreSQL 10 ...
su: Cannot make/remove an entry for the specified session

I even tried compiling without pam, but the result is the same.
any help!

[Moderator edit: added [code] tags to preserve output layout. -Hu]

[limn]

Maybe there is some problem with your postgres user.
See if su logged anything to /var/log/messages.

[tranquilcool]

Maybe there is some problem with your postgres user.
See if su logged anything to /var/log/messages.

nothing in /var/log/messages

[limn]

When it is successfully started on my box, messages has like this:

Code: Select all

Jul 25 06:46:45 gentoo su[10401]: Successful su for postgres by root
Jul 25 06:46:45 gentoo su[10401]: + /dev/pts/0 root:postgres
Jul 25 06:46:45 gentoo su[10401]: pam_unix(su:session): session opened for user postgres by limn(uid=0)
Jul 25 06:46:55 gentoo su[10401]: pam_unix(su:session): session closed for user postgres

dev-db/postgresql-10_beta2 is compiled with "ldap nls pam readline server ssl xml zlib" enabled.

Is sys-apps/shadow-4.5 compiled with pam on your box? Is pam a global use flag?

What happens if you

Code: Select all

su - postgres

[tranquilcool]

When it is successfully started on my box, messages has like this:

Code: Select all

Jul 25 06:46:45 gentoo su[10401]: Successful su for postgres by root
Jul 25 06:46:45 gentoo su[10401]: + /dev/pts/0 root:postgres
Jul 25 06:46:45 gentoo su[10401]: pam_unix(su:session): session opened for user postgres by limn(uid=0)
Jul 25 06:46:55 gentoo su[10401]: pam_unix(su:session): session closed for user postgres

dev-db/postgresql-10_beta2 is compiled with "ldap nls pam readline server ssl xml zlib" enabled.

Is sys-apps/shadow-4.5 compiled with pam on your box? Is pam a global use flag?

What happens if you

Code: Select all

su - postgres

su - postgress asks for password, and i haven't set one.

this is the contents of /etc/pam.d/postgresql-10

Code: Select all

auth    include         system-auth
account include         system-auth
session include         system-auth

contents of /etc/postgresql-10/pg_hba.conf

Code: Select all

# TYPE  DATABASE        USER            ADDRESS                 METHOD
# "local" is for Unix domain socket connections only
local   all             all                                     trust
# IPv4 local connections:
host    all             all             127.0.0.1/32            trust
# IPv6 local connections:
host    all             all             ::1/128                 trust
# Allow replication connections from localhost, by a user with the
# replication privilege.
local   replication     all                                     trust
host    replication     all             127.0.0.1/32            trust
host    replication     all             ::1/128                 trust

this is a line in /var/lib/postgresql/10/data/posmaster.log

Code: Select all

2017-07-28 22:49:45.958 CEST [24261] LOG:  worker process: logical replication launcher (PID 24269) exited with exit code 1

2017-07-28 22:49:45.958 CEST [24261] LOG:  received smart shutdown request
2017-07-28 22:49:46.014 CEST [24264] LOG:  shutting down
2017-07-28 22:49:46.363 CEST [24261] LOG:  database system is shut down

yes shadow is compiled with pam

update:
if i su - postgres as root, i can start the database with the command

Code: Select all

pg_ctl -D /var/lib/postgresql/10/data start

[Moderator edit: added [code] tags to preserve output layout. -Hu]

[limn]

Have you

Code: Select all

eselect postgresql set 10

so that pg_ctl is linked to /usr/lib/postgresql-10/bin/pg_ctl?
What happens if you run

Code: Select all

/usr/lib/postgresql-10/bin/postgres -D /etc/postgresql-10 --data-directory=/var/lib/postgresql/10/data --unix-socket-directories=/run/postgresql

as postgres? That is what the init script runs.

[tranquilcool]

Have you

Code: Select all

eselect postgresql set 10

so that pg_ctl is linked to /usr/lib/postgresql-10/bin/pg_ctl?
What happens if you run

Code: Select all

/usr/lib/postgresql-10/bin/postgres -D /etc/postgresql-10 --data-directory=/var/lib/postgresql/10/data --unix-socket-directories=/run/postgresql

as postgres? That is what the init script runs.

with su - postgres as root, the above commands start the database.
i am thinking maybe it's the init script

[tranquilcool]

yes, the problem was the init script.
i went for the postgresql.init file in the files directory with some corrections
and it worked.

thank you very much.

[limn]

De nada.

[red-0ne]

yes, the problem was the init script.
i went for the postgresql.init file in the files directory with some corrections
and it worked.

I am having the same problem. Could you please indicate how you got it working?

Thank you

[tranquilcool]

yes, the problem was the init script.
i went for the postgresql.init file in the files directory with some corrections
and it worked.

I am having the same problem. Could you please indicate how you got it working?

Thank you

Code: Select all

#!/sbin/openrc-run
# Copyright 1999-2014 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2

extra_started_commands="reload promote"

PG_CTL="/usr/lib64/postgresql-10/bin/pg_ctl"

description="PostgreSQL 10 -- the world's most advanced open source database --
${RC_SERVICE} is a wrapper around pg_ctl with additional administrative checks
and convenience"

get_config() {
    [ -f ${PGDATA%/}/postgresql.conf ] || return 1

    eval echo $(sed -e 's:#.*::' ${PGDATA%/}/postgresql.conf \
        | awk '$1 == "'$1'" { print ($2 == "=" ? $3 : $2) }')
}

depend() {
    use net
    provide postgresql

    if [ "$(get_config log_destination)" = "syslog" ]; then
        use logger
    fi
}

configured_port=$(get_config port)
: ${configured_port:=${PGPORT}}
socket_path=$(get_config unix_socket_directory)
: ${socket_path:=/run/postgresql}

checkconfig() {
    # Check that DATA_DIR has been set
    if [ -z ${DATA_DIR} ] ; then
        eerror "DATA_DIR not set"
        eerror "HINT: Perhaps you need to update /etc/conf.d/postgresql-10"
        return 1
    fi

    # Check that DATA_DIR exists
    if [ ! -d ${DATA_DIR} ] ; then
        eerror "Directory not found: ${DATA_DIR}"
        eerror "HINT: Ensure that DATA_DIR points to the right path."
        eerror "HINT: Or perhaps you need to create the database cluster:"
        eerror "    emerge --config dev-db/postgresql:@SLOT@"
        return 1
    fi

    # Check for the existence of PostgreSQL's config files, and set the
    # proper mode and ownership.
    # Only three files should be checked as potentially other files
    # may be in PGDATA that should not be touched.
    local file
    for file in postgresql pg_hba pg_ident ; do
        file="${PGDATA%/}/${file}.conf"
        if [ -f ${file} ] ; then
            checkpath -f -m 0600 -o postgres:postgres ${file}
        else
            eerror "${file} not found"
            eerror "HINT: mv ${DATA_DIR%/}/*.conf ${PGDATA}"
            return 1
        fi
    done

    # Set the proper permission for the socket path and create it if
    # it doesn't exist.
    checkpath -d -m 1775 -o postgres:postgres ${socket_path}
    if [ -e ${socket_path%/}/.s.PGSQL.${configured_port} ] ; then
        eerror "Socket conflict."
        eerror "A server is already listening on:"
        eerror "    ${socket_path%/}/.s.PGSQL.${configured_port}"
        eerror "HINT: Change PGPORT to listen on a different socket."
        return 1
    fi
}

start() {
    checkconfig || return 1

    ebegin "Starting PostgreSQL"

    rm -f ${DATA_DIR%/}/postmaster.pid

    local extraenv
    local x
    for x in ${PG_EXTRA_ENV} ; do
        extraenv="${extraenv} --env ${x}"
    done

    start-stop-daemon --start \
        --user postgres \
        --env "PGPORT=${configured_port}" \
        ${extraenv} \
        --pidfile ${DATA_DIR%/}/postmaster.pid \
        --exec /usr/lib64/postgresql-10/bin/pg_ctl \
        -- start -s -w -t ${START_TIMEOUT} -l ${DATA_DIR%/}/postmaster.log \
                 -D ${PGDATA} -o "--data-directory=${DATA_DIR} ${PGOPTS}"
    local retval=$?

    if [ $retval -ne 0 ] ; then
        eerror "Check the log for a possible explanation of the above error. The log may be"
        eerror "    ${DATA_DIR%/}/postmaster.log"
        eerror "Or wherever you configured PostgreSQL @SLOT@ log messages to be sent."
        eend $retval
        return $retval
    fi

    eend $retval
}

stop() {
    local seconds=$(( ${NICE_TIMEOUT} + ${RUDE_TIMEOUT} + ${FORCE_TIMEOUT} ))
    ebegin "Stopping PostgreSQL (this can take up to ${seconds} seconds)"

    local retval
    local retries=SIGTERM/${NICE_TIMEOUT}

    if [ "${RUDE_QUIT}" != "NO" ] ; then
        einfo "RUDE_QUIT enabled."
        retries="${retries}/SIGINT/${RUDE_TIMEOUT}"
    fi
    if [ "${FORCE_QUIT}" = "YES" ] ; then
        einfo "FORCE_QUIT enabled."
        ewarn "A recover-run might be executed on next startup."
        retries="${retries}/SIGQUIT/${FORCE_TIMEOUT}"
    fi

    # Loops through nice, rude, and force quit in one go.
    start-stop-daemon --stop \
        --exec /usr/lib64/postgresql-10/bin/pg_ctl \
        --retry ${retries} \
        --pidfile ${DATA_DIR%/}/postmaster.pid

    eend
}

reload() {
    ebegin "Reloading PostgreSQL configuration"
    kill -HUP $(head -n1 ${DATA_DIR%/}/postmaster.pid)
    eend $?
}

this works for me.

[Moderator edit: added [code] tags to preserve output layout. -Hu]