Re: bdb backend - reliable or not?

[Owen DeLong]

This may not be the politically correct thing to say here, but, it is my
experience and YMMV.

BDB is very stable and reliable, once you get the right combination of
configuration figured out.  There's lots of different things that need
to be set up to get BDB stable and many of them are less than obvious
in the installation process.

Running db_recover before EVERY startup is safe and, in my experience,
virtually necessary.  If the shutdown was clean, db_recover will quickly
figure out that it doesn't have to do anything and move on.

Another important thing is to get DB_CONFIG set up correctly.

Here's my LDAP startup script which has been mostly stable:

#!/bin/bash
#
# ldap  This shell script takes care of starting and stopping
#       ldap servers (slapd and slurpd).
#
# chkconfig: - 39 61
# description: LDAP stands for Lightweight Directory Access Protocol, used \
#              for implementing the industry standard directory services.
# processname: slapd
# config: /etc/openldap/slapd.conf
# pidfile: /var/run/slapd.pid

# Source function library.
. /etc/init.d/functions

# Source networking configuration and check that networking is up.
if [ -r /etc/sysconfig/network ] ; then
       . /etc/sysconfig/network
       [ ${NETWORKING} = "no" ] && exit 0
fi

# Source an auxiliary options file if we have one, and pick up OPTIONS,
# SLAPD_OPTIONS, and SLURPD_OPTIONS.
if [ -r /etc/sysconfig/ldap ] ; then
       . /etc/sysconfig/ldap
fi

slapd=/usr/sbin/slapd
slurpd=/usr/sbin/slurpd
slaptest=/usr/sbin/slaptest
[ -x ${slapd} ] || exit 0
[ -x ${slurpd} ] || exit 0

RETVAL=0

function start() {
       # Check for simple-but-common errors.
       user=ldap
       ldapuid=`id -u $user`
       # Unaccessible database files.

               echo "Recovering in ${dbdir}"
               /usr/sbin/slapd_db_recover -h "${dbdir}"

               done
       done
       # Check the configuration file.
       echo -n $"Checking configuration files for $prog: "
       $slaptest -f /etc/openldap/slapd.conf
       RETVAL=$?
       if test $RETVAL -eq 0 ; then
           success
       else
           failure
           exit 1
       fi
       # Start daemons.
       chown -R ldap.ldap /var/lib/ldap
       prog=`basename ${slapd}`
       echo -n $"Starting $prog: "
       if grep -q ^TLS /etc/openldap/slapd.conf ; then

                                                        $OPTIONS $SLAPD_OPTIONS
           RETVAL=$?
       else
           daemon ${slapd} -u ldap -h "ldap:///"; $OPTIONS $SLAPD_OPTIONS
           RETVAL=$?
       fi
       echo
       if [ $RETVAL -eq 0 ]; then
           if grep -q "^replogfile" /etc/openldap/slapd.conf; then
               prog=`basename ${slurpd}`
               echo -n $"Starting $prog: "
               daemon ${slurpd} $OPTIONS $SLURPD_OPTIONS
               RETVAL=$?
               echo
           fi
       fi
       [ $RETVAL -eq 0 ] && touch /var/lock/subsys/ldap
       return $RETVAL
}

function stop() {
       # Stop daemons.
       prog=`basename ${slapd}`
       echo -n $"Stopping $prog: "
       killproc ${slapd}
       RETVAL=$?
       echo
       if [ $RETVAL -eq 0 ]; then
           if grep -q "^replogfile" /etc/openldap/slapd.conf; then
               prog=`basename ${slurpd}`
               echo -n $"Stopping $prog: "
               killproc ${slurpd}
               RETVAL=$?
               echo
           fi
       fi
       [ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/ldap /var/run/slapd.args
       return $RETVAL
}

# See how we were called.
case "$1" in
   start)
       start
       ;;
   stop)
       stop
       ;;
   status)
       status ${slapd}
       if grep -q "^replogfile" /etc/openldap/slapd.conf ; then
           status ${slurpd}
       fi
       ;;
   restart)
       stop
       start
       ;;
   condrestart)
       if [ -f /var/lock/subsys/ldap ] ; then
           stop
           start
       fi
       ;;
   *)
       echo $"Usage: $0 {start|stop|restart|status|condrestart}"
       RETVAL=1
esac

exit $RETVAL







Here is my DB_DONRIG file:

# In Memory Cache Size
set_cachesize   0       8388608         0

# Database Flags


# Log Values
set_lg_regionmax        1048576
set_lg_max              10485760
set_lg_bsize            2097152
set_lg_dir              /var/lib/ldap/bdb-logs

# Temporary file location
set_tmp_dir             /tmp


Hope that helps.

Owen

for dbdir in `grep ^directory /etc/openldap/slapd.conf | sed s,^directory,,` ; do for file in `find ${dbdir}/ -not -uid $ldapuid -and \( -name "*.dbb" -or -name "*.gdbm" -or -name "*.bdb" \)` ; do echo -n $"$file is not owned by \"$user\"" ; warning ; echo daemon ${slapd} -u ldap -h '"ldap://<servername>/ ldaps://<servername>/"' \ #set_flags DB_TXN_NOSYNC # Don't immediately flush txn logs. Dangerous

pgpfiQx2HYIEb.pgp
Description: PGP signature