PostgreSQL /sbin/start-stop-daemon problem - SOLVED

[spottraining]

Hi

I upgrade my baselayout and Postgresql. And now I have problem. I can not start/restart/stop PostgreSQL server with /etc/init.d/postgresql ..... command. I getting this error:

one init.d # /etc/init.d/postgresql stop
* Stopping PostgreSQL ...
/sbin/start-stop-daemon: invalid schedule item (must be [-]<signal-name>, -<signal-number>, <timeout> or `forever'
Try `/sbin/start-stop-daemon --help' for more information. [ !! ]

I have found, that some bug was with baselayout - but version 1.13. I have now baselayout-1.12.6, PostgreSQL is postgresql-8.0.9-r1.

Other init.d scripts working....

How to fix this problem? I can only start PostgreSQL server, when I do /etc/init.d/postgresql zap and then start. PostgreSQL server with databases works.

[PaulBredbury]

The problem may be with the ENV_* variables in /etc/conf.d/postgresql

[spottraining]

The problem may be with the ENV_* variables in /etc/conf.d/postgresql

and how to fix that?
My /etc/conf.d/postgresql file:

Code: Select all

# PostgreSQL's Database Directory
PGDATA=/var/lib/postgresql/data

# Logfile path: (NOTE: This must be uid/gid owned by the value of $PGUSER!)
PGLOG=/var/lib/postgresql/data/postgresql.log

# Run the PostgreSQL user as:
PGUSER=postgres

# Extra options to run postmaster with.
# If you want to enable TCP/IP for PostgreSQL, add -i to the following:
# PGOPTS="-N 1024 -B 2048 -i"
PGOPTS=""

[PaulBredbury]

It's not that then, if there are no ENV_* variables defined. I don't see what else it can be - my start-stop line in /etc/init.d/postgresql (with dev-db/postgresql-8.1.5-r1) is:

Code: Select all

    start-stop-daemon --start \
        --pidfile "${PGDATA}/postmaster.pid" \
        --chuid ${PGUSER}:${PGGROUP} \
        --exec /usr/bin/postmaster \
        ${env_vars} -- \
        -D "${PGDATA}" \
        --silent-mode=true \
        ${PGOPTS}

[spottraining]

My lines are:

Code: Select all

 local env_vars
        for var in ${!ENV_*} ; do env_vars="${env_vars} -e ${!var}" ; done
        start-stop-daemon --start \
                --pidfile "${PGDATA}/postmaster.pid" \
                --chuid ${PGUSER}:${PGGROUP} \
                --exec /usr/bin/postmaster \
                ${env_vars} -- \
                -D "${PGDATA}" \
                --silent-mode=true \
                ${PGOPTS}

        eend $?
}

[spottraining]

up

[jfb3]

I had the same problem stopping postgresql until I realized I hadn't run etc-update. The resolution was that I needed to update the /etc/conf.d/postgresl file to add the following:

Code: Select all

# SERVER SHUTDOWN:
# The server will receive 3 signals in the worst case:
# 1. SIGTERM
#  This signals the server to ignore new connections and to
#  wait for all clients to end their transactions before shutting down.
#  Use WAIT_FOR_DISCONNECT to control how much time the clients
#  should have until the next signal is being sent.
# 2. SIGINT
#  Tell the server to forcefully disconnect all clients.
#  Terminating a client results in a rollback of the open transactions for this client.
#  Use WAIT_FOR_CLEANUP to determine how much time the server has
#  for cleanup. (Set it to "forever" if you want to wait forever.)
# 3. SIGQUIT
#  This will terminate the server immediately and results in a recovery run for the next start.

# Wait for clients to disconnect (seconds or "forever")
WAIT_FOR_DISCONNECT=10

# Time the server has to clean up (seconds or "forever")
WAIT_FOR_CLEANUP=10

# If you have to export environment variables for the database process,
# this can be done here.
# Just define a new variable ENV_identifier="ENV_VAR_NAME=env_var_content",
# ENV_VAR_NAME will then be exported to the postgresql's environment.
#
# Example:
#   ENV_plr="R_HOME=/usr/lib/R"
# This exports a variable R_HOME with the content "/usr/lib/R"

Now I can stop postgresql just fine.