On Wed, Apr 18, 2007 at 12:47:10AM +0200, Lars Marowsky-Bree wrote:
> On 2007-04-17T19:53:43, [EMAIL PROTECTED] wrote:
>
> > Attached is an eDirectory OCF RA, which has now seen some real-life
> > cluster testing, as well as been reviewed by various people.
> >
> > I thas been tested with up to 3 eDir instances on a 4-node cluster.
> >
> > It is specifically designed to handle eDirectory 8.8 (don't try on
> > earlier versions, it won't work), including the new "multiple
> > instances" feature.
> >
> > Any and all feedback (positive, negative, or otherwise) welcome.
>
> Hi Yan, thanks for this version!
>
> I'm adding it locally to my tree, but before pushing it out, I'd like to
> discuss some points in the script. See the attached patch as a
> suggestion for incorporating this feedback:
>
> > #!/bin/sh
> > #
> > # eDirectory Resource Agent (RA) for Heartbeat.
> > # This script is only compatible with eDirectory 8.8 and later
>
> If it's only compatible with said version, maybe the script should check
> that somewhere?
>
> > #
> > # Copyright (c) 2007 SUSE LINUX AG, Yan Fitterer
> > # All Rights Reserved.
>
> You mean Novell Inc, right?
>
> > #
> > # This program is free software; you can redistribute it and/or modify
> > # it under the terms of version 2 of the GNU General Public License as
> > # published by the Free Software Foundation.
> > #
> > # This program is distributed in the hope that it would be useful, but
> > # WITHOUT ANY WARRANTY; without even the implied warranty of
> > # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
> > #
> > # Further, this software is distributed without any warranty that it is
> > # free of the rightful claim of any third person regarding infringement
> > # or the like. Any license provided herein, whether implied or
> > # otherwise, applies only to this software file. Patent licenses, if
> > # any, provided herein do not apply to combinations of this program with
> > # other software, or any other product whatsoever.
> > #
> > # You should have received a copy of the GNU General Public License
> > # along with this program; if not, write the Free Software Foundation,
> > # Inc., 59 Temple Place - Suite 330, Boston MA 02111-1307, USA.
> > #
> > #
> > # OCF parameters:
> > # OCF_RESKEY_eDir_config_file - full filename to instance
> > configuration file
> > # OCF_RESKEY_eDir_monitor_ldap - Should we monitor LDAP (0/1 - 1 is
> > true)
> > # OCF_RESKEY_eDir_monitor_idm - Should we monitor IDM (0/1 - 1 is
> > true)
> > # OCF_RESKEY_eDir_jvm_initial_heap - Value of the DHOST_INITIAL_HEAP java
> > env var
> > # OCF_RESKEY_eDir_jvm_max_heap - Value of the DHOST_MAX_HEAP java env
> > var
> > # OCF_RESKEY_eDir_jvm_options - Value of the DHOST_OPTIONS java env
> > var
> > ###############################################################################
> >
> > #######################################################################
> > # Initialization:
> >
> > . /usr/lib/heartbeat/ocf-shellfuncs
>
> I'm substituting the init path here.
>
> > #######################################################################
> >
> > usage() {
> > ME=$(basename "$0")
> > cat <<-EOFA
> >
> > usage: $ME start|stop|status|monitor|methods|validate-all
> >
> > $ME manages an eDirectory instance as an HA resource.
> >
> > The 'start' operation starts the instance.
> > The 'stop' operation stops the instance.
> > The 'status' opneration reports if the instance is running.
> > The 'monitor' operation reports if the instance is running, and runs
> > additional checks.
> > The 'methods' operation reports on the methods $0 supports.
> > The 'validate-all' operation checks the validity of the arguments
> > (environment variables).
> > EOFA
> > return $OCF_ERR_ARGS
> > }
> >
> > eDir_meta_data() {
> > cat <<-EOFB
> > <?xml version="1.0"?>
> > <!DOCTYPE resource-agent SYSTEM "ra-api-1.dtd">
> > <resource-agent name="eDirectory" version="0.9">
> > <version>1.0</version>
> >
> > <longdesc lang="en">
> > Resource script for managing an eDirectory instance. Manages a single
> > instance
> > of eDirectory as an HA resource. The "multiple instances" feature or
> > eDirectory has been added in version 8.8. This script will not work for any
> > version of eDirectory prior to 8.8. This RA can be used to load multiple
> > eDirectory instances on the same host.
> >
> > Setting a high heartbeat timeout value is _very_ _strongly_ recommended.
> > eDir
> > with IDM can take in excess of 10 minutes to start. If heartbeat times out
> > before eDir has had a chance to start properly, mayhem _WILL ENSUE_.
> >
> > The LDAP module seems to be one of the very last to start. So this script
> > will
> > take even longer to start on installations with IDM and LDAP if the
> > monitoring
> > of IDM and/or LDAP is enabled, as the start command will wait for IDM and
> > LDAP
> > to be available.
> > </longdesc>
> > <shortdesc lang="en">eDirectory resource agent</shortdesc>
> > <parameters>
> > <parameter name="eDir_config_file" unique="1" required="1">
> > <longdesc lang="en">
> > Path to configuration file for eDirectory instance.
> > </longdesc>
> > <shortdesc lang="en">eDir config file</shortdesc>
> > <content type="string" default="/etc/opt/novell/eDirectory/conf/nds.conf" />
> > </parameter>
> > <parameter name="eDir_monitor_ldap" unique="0" required="0">
> > <longdesc lang="en">
> > Should we monitor if LDAP is running for the eDirectory instance?
> > </longdesc>
> > <shortdesc lang="en">eDir monitor ldap</shortdesc>
> > <content type="boolean" default="0" />
> > </parameter>
> > <parameter name="eDir_monitor_idm" unique="0" required="0">
> > <longdesc lang="en">
> > Should we monitor if IDM is running for the eDirectory instance?
> > </longdesc>
> > <shortdesc lang="en">eDir monitor IDM</shortdesc>
> > <content type="boolean" default="0" />
> > </parameter>
> > <parameter name="eDir_startup_timeout" unique="0" required="0">
> > <longdesc lang="en">
> > How long (in seconds) do we wait for eDir to start this instance (give
> > _plenty_ of time here...)?
> > </longdesc>
> > <shortdesc lang="en">eDir startup timeout</shortdesc>
> > <content type="boolean" default="0" />
> > </parameter>
> > <parameter name="eDir_jvm_initial_heap" unique="0" required="0">
> > <longdesc lang="en">
> > Value for the DHOST_INITIAL_HEAP java environment variable. If unset, java
> > defaults will be used.
> > </longdesc>
> > <shortdesc lang="en">DHOST_INITIAL_HEAP value</shortdesc>
> > <content type="integer" default="" />
> > </parameter>
> > <parameter name="eDir_jvm_max_heap" unique="0" required="0">
> > <longdesc lang="en">
> > Value for the DHOST_MAX_HEAP java environment variable. If unset, java
> > defaults will be used.
> > </longdesc>
> > <shortdesc lang="en">DHOST_MAX_HEAP value</shortdesc>
> > <content type="integer" default="" />
> > </parameter>
> > <parameter name="eDir_jvm_options" unique="0" required="0">
> > <longdesc lang="en">
> > Value for the DHOST_OPTIONS java environment variable. If unset, original
> > values will be used.
> > </longdesc>
> > <shortdesc lang="en">DHOST_OPTIONS value</shortdesc>
> > <content type="string" default="" />
> > </parameter>
> > </parameters>
> >
> > <actions>
> > <action name="start" timeout="250" />
> > <action name="stop" timeout="250" />
>
> Are these timeouts sufficient? When you mention in excess of 10 minutes,
> this is just roughly 4 minutes ...
>
> > <action name="monitor" timeout="60" interval="30" />
> > <action name="meta-data" timeout="5" />
> > <action name="methods" timeout="5" />
> > <action name="validate-all" timeout="5">
> > </actions>
> > </resource-agent>
> > EOFB
> > return $OCF_SUCCESS
> > }
> >
> >
> > #
> > # methods: What methods/operations do we support?
> > #
> >
> > eDir_methods() {
> > cat <<-EOFC
> > start
> > stop
> > status
> > monitor
> > methods
> > meta-data
> > validate-all
> > EOFC
> > }
>
> Do we need the methods call? It's totally useless, implied by usage and
> not used anywhere ... I wonder where that one comes from?
>
> > #
> > # eDir_start: Start eDirectory instance
> > #
> >
> > eDir_start() {
> > if eDir_status ; then
> > ocf_log info "eDirectory is already running ($NDSCONF)."
> > return $OCF_SUCCESS
> > fi
> >
> > # Start eDirectory instance
> > if [ "$OCF_RESKEY_eDir_jvm_initial_heap" ]; then
> > DHOST_JVM_INITIAL_HEAP=$OCF_RESKEY_eDir_jvm_initial_heap
> > export DHOST_JVM_INITIAL_HEAP
> > fi
> > if [ "$OCF_RESKEY_eDir_jvm_max_heap" ]; then
> > DHOST_JVM_MAX_HEAP=$OCF_RESKEY_eDir_jvm_max_heap
> > export DHOST_JVM_MAX_HEAP
> > fi
> > if [ "$OCF_RESKEY_eDir_jvm_options" ]; then
> > DHOST_JVM_OPTIONS=$OCF_RESKEY_eDir_jvm_options
> > export DHOST_JVM_OPTIONS
> > fi
>
> Cleaner to add a -n to all of these.
>
> > $NDSMANAGE start --config-file "$NDSCONF" > /dev/null 2>&1
> > if [ $? -eq 0 ]; then
> > ocf_log info "eDir start command sent for $NDSCONF."
> > else
> > echo "ERROR: Can't start eDirectory for $NDSCONF."
> > return $OCF_ERR_GENERIC
> > fi
> >
> > if ! eDir_monitor ; then
> > # Apparently, LDAP will only start after all other services
> > # Startup time can be in excess of 10 minutes.
> > # Leave a very long heartbeat timeout on the start action
> > # We're relying on heartbeat to bail us out...
> > sleep 10
> > while [ "$((CNT++))" ]; do
> > if eDir_monitor ; then
> > ocf_log info "eDirectory start verified after $CNT retries
> > for $NDSCONF."
> > break
> > fi
> > sleep 10
> > done
> > fi
> >
> > ocf_log info "eDirectory start verified for $NDSCONF."
> > return $OCF_SUCCESS
> > }
>
> Ouch. But probably needed. I've clarified this code a bit.
>
> >
> > #
> > # eDir_stop: Stop eDirectory instance
> > # This action is written in such a way that even when run
> > # on a node were things are broken (no binaries, no config
> > # etc...) it will try to stop any running ndsd processes
> > # and report success if none are running.
> > #
> >
> > eDir_stop() {
> > if ! eDir_status ; then
> > return $OCF_SUCCESS
> > fi
> > declare -a procs
Does /bin/sh support arrays? I think that so far all resource
agents were supported by the most common denominator which is
Bourne shell.
This can be easily handled by a string containing a list of space
separated pids.
> > procs=($(pgrep '^ndsd$'))
>
> That's bash specific syntax, so you likely want #!/bin/bash at the top
> and not /bin/sh.
See above.
> > while [ -z ] ; do
>
> That's meant to be a -z $procs, right?
>
> And, this seems to not handle several ndsd instances active. Can there
> ever be only one per node, even though one can configure several
> resources?
>
> > if [ -r "$NDSCONF" -a -x $NDSMANAGE ] ; then
> > $NDSMANAGE stop --config-file "$NDSCONF" >/dev/null 2>&1
> > if eDir_status ; then
> > # eDir failed to stop.
> > ocf_log err "eDirectory instance failed to stop for $NDSCONF"
> > else
> > ocf_log info "eDirectory stop verified for $NDSCONF."
> > return $OCF_SUCCESS
> > fi
> > else
> > ocf_log err "ndsmanage binary or config file missing ($NDSCONF)"
> > if [ [EMAIL PROTECTED] -eq 1 ] ; then
> > ocf_log err "Single ndsd process found (PID ${procs[0]}).
> > Terminating."
> > kill -15 ${procs[0]}
> > sleep 20
>
> 20s seems like a magic number. Is that sufficient? And, should you maybe
> escalate to -9 if everything fails?
I think that we once agreed that the LRM should handle runaway
processes, i.e. we send the process a signal, perhaps several,
it's up to the implementor and depends on the service, and wait
_forever_ (probably in a loop with small sleeps) until the process
exits. A user defined timeout in the CIB should say when enough is
enough. There are, however, many RAs with magical timeouts.
> > else
> > ocf_log err "Too many ndsd processes found to kill safely (
> > [EMAIL PROTECTED] ). STOP action failed."
> > return $OCF_ERR_GENERIC
> > fi
>
> And why is that a problem if there's several instances ...?
>
> > if [ -z $(ps -p ${procs[0]} -o pid=) ] ; then
>
> kill -0 on the pid is a lot simpler.
Right. BTW, did you know that 'kill -0' returns success in case
the process is a zombie? That's where 'kill' method differs from
'ps | grep'.
> > ocf_log info "eDirectory stop verified for $NDSCONF."
> > return $OCF_SUCCESS
> > fi
> > fi
> >
> > done
> > }
> >
> > #
> > # eDir_status: is eDirectory instance up (do simple ndsstat check)?
> > #
> >
> > eDir_status() {
> > if [ -r "$NDSCONF" -a -x $NDSSTAT ] ; then
> > $NDSSTAT --config-file "$NDSCONF" >/dev/null 2>&1
> > if [ $? -eq 0 ] ; then
> > return 0
> > else
> > return 1
> > fi
> > else
> > # eDir isn't installed, or this RA not configured correctly
> > # Try to see if ndsd is running anyway.
> > if [ -n "$(pgrep '^ndsd$')" ] ; then
> > return 0
> > else
> > return 1
> > fi
>
> This will return success for every ndsd instance, even if it's not the
> one the RA is supposed to manage. And the stop operation will mishandle
> that case as well. I'm not convinced this is a good approach.
>
> > fi
> > }
> >
> >
> > #
> > # eDir_monitor: Do several checks to ensure that eDirectory is fully
> > functional
> > # LDAP and IDM checks are only done if reqested.
> > #
> >
> > eDir_monitor() {
> > if ! eDir_status ; then
> > ocf_log info "eDirectory instance is down ($NDSCONF)"
> > return $OCF_NOT_RUNNING
> > fi
> >
> > # Monitor IDM first, as it will start before LDAP
> > if [ $MONITOR_IDM -eq 1 ]; then
> > RET=$($NDSTRACE --config-file "$NDSCONF" -c modules | egrep -i
> > '^vrdim.*Running' | awk '{print $1}')
> > if [ "$RET" != "vrdim" ]; then
> > ocf_log err "eDirectory IDM engine isn't running ($NDSCONF)."
> > return $OCF_ERR_GENERIC
> > fi
> > fi
> > if [ $MONITOR_LDAP -eq 1 ] ; then
> > $NDSNLDAP -c --config-file "$NDSCONF" >/dev/null 2>&1
> > if [ $? -ne 0 ]; then
> > ocf_log err "eDirectory LDAP server isn't running ($NDSCONF)."
> > return $OCF_ERR_GENERIC
> > fi
> > fi
> >
> > ocf_log debug "eDirectory monitor success ($NDSCONF)"
> > return $OCF_SUCCESS
> > }
> >
> > #
> > # eDir_validate: Validate environment
> > #
> >
> > eDir_validate() {
> >
> > declare VALID=$OCF_SUCCESS
> >
> > # Script must be run as root
> > if ! ocf_is_root ; then
> > ocf_log err "$0 must be run as root"
> > VALID=$OCF_ERR_GENERIC
>
> "VALID=ERROR"? You got to be kidding me ;-) I'm suggesting to change
> this to "rc".
>
> > fi
> >
> > # ndsmanage must be available and runnable
> > if [ ! -x $NDSMANAGE ] ; then
> > ocf_log err "Cannot run $NDSMANAGE"
> > VALID=$OCF_ERR_INSTALLED
> > fi
> >
> > # ndsstat must be available and runnable
> > if [ ! -x $NDSSTAT ]; then
> > ocf_log err "Cannot run $NDSSTAT"
> > VALID=$OCF_ERR_INSTALLED
> > fi
> >
> > # Config file must be readable
> > if [ ! -r "$NDSCONF" ] ; then
> > ocf_log err "eDirectory configuration file [$NDSCONF] is not
> > readable"
> > VALID=$OCF_ERR_ARGS
> > fi
> >
> > # monitor_ldap must be unambiguously resolvable to a truth value
> > MONITOR_LDAP=$(echo "$MONITOR_LDAP" | tr [A-Z] [a-z])
> > case "$MONITOR_LDAP" in
> > yes|true|1)
> > MONITOR_LDAP=1;;
> > no|false|0)
> > MONITOR_LDAP=0;;
> > *)
> > ocf_log err "Configuration parameter eDir_monitor_ldap has
> > invalid value [$MONITOR_LDAP]"
> > VALID=$OCF_ERR_ARGS;;
> > esac
> >
> > # monitor_idm must be unambiguously resolvable to a truth value
> > MONITOR_IDM=$(echo "$MONITOR_IDM" | tr [A-Z] [a-z])
> > case "$MONITOR_IDM" in
> > yes|true|1)
> > MONITOR_IDM=1;;
> > no|false|0)
> > MONITOR_IDM=0;;
> > *)
> > ocf_log err "Configuration parameter eDir_monitor_idm has
> > invalid value [$MONITOR_IDM]"
> > VALID=$OCF_ERR_ARGS;;
> > esac
> >
> > # eDir_jvm_initial_heap must be blank or numeric
> > if [ -n "$OCF_RESKEY_eDir_jvm_initial_heap" ] ; then
> > if ! ocf_is_decimal "$OCF_RESKEY_eDir_jvm_initial_heap" ; then
> > ocf_log err "Configuration parameter eDir_jvm_initial_heap has
> > invalid" \
> > "value [$OCF_RESKEY_eDir_jvm_initial_heap]"
> > VALID=$OCF_ERR_ARGS
> > fi
> > fi
> >
> > # eDir_jvm_max_heap must be blank or numeric
> > if [ -n "$OCF_RESKEY_eDir_jvm_max_heap" ] ; then
> > if ! ocf_is_decimal "$OCF_RESKEY_eDir_jvm_max_heap" ; then
> > ocf_log err "Configuration parameter eDir_jvm_max_heap has
> > invalid" \
> > "value [$OCF_RESKEY_eDir_jvm_max_heap]"
> > VALID=$OCF_ERR_ARGS
> > fi
> > fi
> > if [ $VALID -ne $OCF_SUCCESS ] ; then
> > ocf_log err "Invalid environment"
> > fi
> > return $VALID
> > }
> >
> > #
> > # Start of main logic
> > #
> >
> > ocf_log debug "$0 started with arguments \"[EMAIL PROTECTED]""
> >
> > NDSBASE=/opt/novell/eDirectory
> > NDSNLDAP=$NDSBASE/sbin/nldap
> > NDSMANAGE=$NDSBASE/bin/ndsmanage
> > NDSSTAT=$NDSBASE/bin/ndsstat
> > NDSTRACE=$NDSBASE/bin/ndstrace
>
> Ohhh. Magic path names! Are these guaranteed to be right? Should they be
> configurable?
Yes, most definitely they should. I can bet that in the next
release it will move to /usr/opt :) Never ever make such assumptions.
> > NDSCONF=${OCF_RESKEY_eDir_config_file:-/etc/opt/novell/eDirectory/conf/nds.conf}
> > MONITOR_LDAP=${OCF_RESKEY_eDir_monitor_ldap:-0}
> > MONITOR_IDM=${OCF_RESKEY_eDir_monitor_idm:-0}
> >
> >
> > # What kind of method was invoked?
> > case "$1" in
> > validate-all) eDir_validate; exit $?;;
> > methods) eDir_methods; exit $OCF_SUCCESS;;
> > meta-data) eDir_meta_data; exit $OCF_SUCCESS;;
> > stop) eDir_stop; exit $?;;
> > status) if eDir_status ; then
> > ocf_log info "eDirectory instance is up ($NDSCONF)"
> > exit $OCF_SUCCESS
> > else
> > ocf_log info "eDirectory instance is down ($NDSCONF)"
> > exit $OCF_NOT_RUNNING
> > fi;;
> > start) : skip;;
> > monitor) : skip;;
> > *) ocf_log err "Invalid argument [$1]"
> > usage; exit $OCF_ERR_ARGS;;
> > esac
> >
> > # From now on we must have a valid environment to continue.
> > # stop goes in the list above as it should ideally be able to
> > # clean up after a start that failed due to bad args
> >
> > eDir_validate
> > RC=$?
> > if [ $RC != $OCF_SUCCESS ]; then
> > exit $RC
> > fi
> >
> > case "$1" in
> > start) eDir_start;;
> > monitor) eDir_monitor;;
> > esac
> >
> > exit $?
>
> Please let me know your thoughts.
>
>
> Sincerely,
> Lars
>
> --
> Teamlead Kernel, SuSE Labs, Research and Development
> SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
> "Experience is the name everyone gives to their mistakes." -- Oscar Wilde
>
> changeset: 10478:fb7399480c5b
> user: Lars Marowsky-Bree <[EMAIL PROTECTED]>
> date: Wed Apr 18 00:44:33 2007 +0200
> summary: RA: eDirectory: Some cleanups.
>
> diff -r 8ecbff996f7e -r fb7399480c5b resources/OCF/eDirectory.in
> --- a/resources/OCF/eDirectory.in Wed Apr 18 00:21:15 2007 +0200
> +++ b/resources/OCF/eDirectory.in Wed Apr 18 00:44:33 2007 +0200
> @@ -1,9 +1,9 @@
> -#!/bin/sh
> +#!/bin/bash
> #
> # eDirectory Resource Agent (RA) for Heartbeat.
> # This script is only compatible with eDirectory 8.8 and later
> #
> -# Copyright (c) 2007 SUSE LINUX AG, Yan Fitterer
> +# Copyright (c) 2007 Novell Inc, Yan Fitterer
> # All Rights Reserved.
> #
> # This program is free software; you can redistribute it and/or modify
> @@ -38,7 +38,7 @@
> #######################################################################
> # Initialization:
>
> -. /usr/lib/heartbeat/ocf-shellfuncs
> +. @hb_libdir@/ocf-shellfuncs
>
> #######################################################################
>
> @@ -74,7 +74,7 @@ version of eDirectory prior to 8.8. This
> version of eDirectory prior to 8.8. This RA can be used to load multiple
> eDirectory instances on the same host.
>
> -Setting a high heartbeat timeout value is _very_ _strongly_ recommended. eDir
> +Setting a high action timeout value is _very_ _strongly_ recommended. eDir
> with IDM can take in excess of 10 minutes to start. If heartbeat times out
> before eDir has had a chance to start properly, mayhem _WILL ENSUE_.
>
> @@ -92,42 +92,42 @@ Path to configuration file for eDirector
> <shortdesc lang="en">eDir config file</shortdesc>
> <content type="string" default="/etc/opt/novell/eDirectory/conf/nds.conf" />
> </parameter>
> -<parameter name="eDir_monitor_ldap" unique="0" required="0">
> +<parameter name="eDir_monitor_ldap" required="0">
> <longdesc lang="en">
> Should we monitor if LDAP is running for the eDirectory instance?
> </longdesc>
> <shortdesc lang="en">eDir monitor ldap</shortdesc>
> <content type="boolean" default="0" />
> </parameter>
> -<parameter name="eDir_monitor_idm" unique="0" required="0">
> +<parameter name="eDir_monitor_idm" required="0">
> <longdesc lang="en">
> Should we monitor if IDM is running for the eDirectory instance?
> </longdesc>
> <shortdesc lang="en">eDir monitor IDM</shortdesc>
> <content type="boolean" default="0" />
> </parameter>
> -<parameter name="eDir_startup_timeout" unique="0" required="0">
> +<parameter name="eDir_startup_timeout" required="0">
> <longdesc lang="en">
> How long (in seconds) do we wait for eDir to start this instance (give
> _plenty_ of time here...)?
> </longdesc>
> <shortdesc lang="en">eDir startup timeout</shortdesc>
> <content type="boolean" default="0" />
> </parameter>
> -<parameter name="eDir_jvm_initial_heap" unique="0" required="0">
> +<parameter name="eDir_jvm_initial_heap" required="0">
> <longdesc lang="en">
> Value for the DHOST_INITIAL_HEAP java environment variable. If unset, java
> defaults will be used.
> </longdesc>
> <shortdesc lang="en">DHOST_INITIAL_HEAP value</shortdesc>
> <content type="integer" default="" />
> </parameter>
> -<parameter name="eDir_jvm_max_heap" unique="0" required="0">
> +<parameter name="eDir_jvm_max_heap" required="0">
> <longdesc lang="en">
> Value for the DHOST_MAX_HEAP java environment variable. If unset, java
> defaults will be used.
> </longdesc>
> <shortdesc lang="en">DHOST_MAX_HEAP value</shortdesc>
> <content type="integer" default="" />
> </parameter>
> -<parameter name="eDir_jvm_options" unique="0" required="0">
> +<parameter name="eDir_jvm_options" required="0">
> <longdesc lang="en">
> Value for the DHOST_OPTIONS java environment variable. If unset, original
> values will be used.
> </longdesc>
> @@ -177,15 +177,15 @@ eDir_start() {
> fi
>
> # Start eDirectory instance
> - if [ "$OCF_RESKEY_eDir_jvm_initial_heap" ]; then
> + if [ -n "$OCF_RESKEY_eDir_jvm_initial_heap" ]; then
> DHOST_JVM_INITIAL_HEAP=$OCF_RESKEY_eDir_jvm_initial_heap
> export DHOST_JVM_INITIAL_HEAP
> fi
> - if [ "$OCF_RESKEY_eDir_jvm_max_heap" ]; then
> + if [ -n "$OCF_RESKEY_eDir_jvm_max_heap" ]; then
> DHOST_JVM_MAX_HEAP=$OCF_RESKEY_eDir_jvm_max_heap
> export DHOST_JVM_MAX_HEAP
> fi
> - if [ "$OCF_RESKEY_eDir_jvm_options" ]; then
> + if [ -n "$OCF_RESKEY_eDir_jvm_options" ]; then
> DHOST_JVM_OPTIONS=$OCF_RESKEY_eDir_jvm_options
> export DHOST_JVM_OPTIONS
> fi
> @@ -198,22 +198,19 @@ eDir_start() {
> return $OCF_ERR_GENERIC
> fi
>
> - if ! eDir_monitor ; then
> + CNT=0
> + while ! eDir_monitor ; do
> # Apparently, LDAP will only start after all other services
> # Startup time can be in excess of 10 minutes.
> # Leave a very long heartbeat timeout on the start action
> # We're relying on heartbeat to bail us out...
> + CNT=$[CNT+1}
> + ocf_log info "eDirectory start waiting for ${CNT}th retry for
> $NDSCONF."
> sleep 10
> - while [ "$((CNT++))" ]; do
> - if eDir_monitor ; then
> - ocf_log info "eDirectory start verified after $CNT retries
> for $NDSCONF."
> - break
> - fi
> - sleep 10
> - done
> - fi
> -
> + done
> +
> ocf_log info "eDirectory start verified for $NDSCONF."
> +
> return $OCF_SUCCESS
> }
>
> @@ -231,7 +228,7 @@ eDir_stop() {
> fi
> declare -a procs
> procs=($(pgrep '^ndsd$'))
> - while [ -z ] ; do
> + while [ -z "[EMAIL PROTECTED]" ] ; do
> if [ -r "$NDSCONF" -a -x $NDSMANAGE ] ; then
> $NDSMANAGE stop --config-file "$NDSCONF" >/dev/null 2>&1
> if eDir_status ; then
> @@ -251,7 +248,7 @@ eDir_stop() {
> ocf_log err "Too many ndsd processes found to kill safely (
> [EMAIL PROTECTED] ). STOP action failed."
> return $OCF_ERR_GENERIC
> fi
> - if [ -z $(ps -p ${procs[0]} -o pid=) ] ; then
> + if kill -0 ${procs[0]} >/dev/null 2>&1 ; then
> ocf_log info "eDirectory stop verified for $NDSCONF."
> return $OCF_SUCCESS
> fi
> @@ -321,30 +318,30 @@ eDir_monitor() {
>
> eDir_validate() {
>
> - declare VALID=$OCF_SUCCESS
> + declare rc=$OCF_SUCCESS
>
> # Script must be run as root
> if ! ocf_is_root ; then
> ocf_log err "$0 must be run as root"
> - VALID=$OCF_ERR_GENERIC
> + rc=$OCF_ERR_GENERIC
> fi
>
> # ndsmanage must be available and runnable
> if [ ! -x $NDSMANAGE ] ; then
> ocf_log err "Cannot run $NDSMANAGE"
> - VALID=$OCF_ERR_INSTALLED
> + rc=$OCF_ERR_INSTALLED
> fi
>
> # ndsstat must be available and runnable
> if [ ! -x $NDSSTAT ]; then
> ocf_log err "Cannot run $NDSSTAT"
> - VALID=$OCF_ERR_INSTALLED
> + rc=$OCF_ERR_INSTALLED
> fi
>
> # Config file must be readable
> if [ ! -r "$NDSCONF" ] ; then
> ocf_log err "eDirectory configuration file [$NDSCONF] is not
> readable"
> - VALID=$OCF_ERR_ARGS
> + rc=$OCF_ERR_ARGS
> fi
>
> # monitor_ldap must be unambiguously resolvable to a truth value
> @@ -356,7 +353,7 @@ eDir_validate() {
> MONITOR_LDAP=0;;
> *)
> ocf_log err "Configuration parameter eDir_monitor_ldap has
> invalid value [$MONITOR_LDAP]"
> - VALID=$OCF_ERR_ARGS;;
> + rc=$OCF_ERR_ARGS;;
> esac
>
> # monitor_idm must be unambiguously resolvable to a truth value
> @@ -368,7 +365,7 @@ eDir_validate() {
> MONITOR_IDM=0;;
> *)
> ocf_log err "Configuration parameter eDir_monitor_idm has
> invalid value [$MONITOR_IDM]"
> - VALID=$OCF_ERR_ARGS;;
> + rc=$OCF_ERR_ARGS;;
> esac
>
> # eDir_jvm_initial_heap must be blank or numeric
> @@ -376,7 +373,7 @@ eDir_validate() {
> if ! ocf_is_decimal "$OCF_RESKEY_eDir_jvm_initial_heap" ; then
> ocf_log err "Configuration parameter eDir_jvm_initial_heap has
> invalid" \
> "value [$OCF_RESKEY_eDir_jvm_initial_heap]"
> - VALID=$OCF_ERR_ARGS
> + rc=$OCF_ERR_ARGS
> fi
> fi
>
> @@ -385,13 +382,13 @@ eDir_validate() {
> if ! ocf_is_decimal "$OCF_RESKEY_eDir_jvm_max_heap" ; then
> ocf_log err "Configuration parameter eDir_jvm_max_heap has
> invalid" \
> "value [$OCF_RESKEY_eDir_jvm_max_heap]"
> - VALID=$OCF_ERR_ARGS
> + rc=$OCF_ERR_ARGS
> fi
> fi
> if [ $VALID -ne $OCF_SUCCESS ] ; then
> ocf_log err "Invalid environment"
> fi
> - return $VALID
> + return $rc
> }
>
> #
> @@ -435,7 +432,7 @@ esac
>
> eDir_validate
> RC=$?
> -if [ $RC != $OCF_SUCCESS ]; then
> +if [ $RC -ne $OCF_SUCCESS ]; then
> exit $RC
> fi
>
>
> _______________________________________________________
> Linux-HA-Dev: [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
--
Dejan
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
