[Pacemaker] ocf_log

[Steve Kwee]

Steve Kwee <skwee at ...> writes:

> 
> Andrew Beekhof <beekhof at ...> writes:
> 
> > 
> > On Fri, Apr 17, 2009 at 13:28, Steve Kwee <skwee at ...> wrote:
> > >
> > > In my /etc/ha.cf I now have:
> ....
> > Opps - this makes no difference for openais clusters (I missed that
> > part of your original email).
> > 
> > find this part of openais.conf and change debug: to "on"
> > 
> > logging {
> > 	debug: off
> > 	fileline: off
> > 	to_syslog: yes
> > 	to_stderr: no
> > 	syslog_facility: daemon
> > 	timestamp: on
> > }
> > 
> ....
> 
Thanks again for your answers
Here is my summary on the subject of
logging in an openAIS / pacemaker stack

1. Logging for heartbeat processes
The logging for openAIS is configured in /etc/ais/openais.conf
example:

 logging {
 	fileline: off
 	to_syslog: yes
 	to_stderr: no
 	syslog_facility: daemon
 	# debug: on
 	debug: off
 	timestamp: on
 }
for details see the manpage openais.conf

The debug flag is the default setting for the child processes 
(crmd, lrmd, mgmtd etc)
The individual debug level of each daemon can dynamically be changed by
sending the heartbeat process SIGUSR1 and SIGUSR2 signals. SIGUSR1 raises the
debug level, and SIGUSR2 lowers it. A sample debug directive is shown below.
When the heartbeat subprocess dies, it will be recreated by aisexec with 
the default setting. So debugging ca be turned on dynamically for individual 
child processes even if the setting in openais.conf is "off"

The config file openais.conf is read on startup. There is no known way 
to reload it but to restart the openAIS daemon

2. Logging for OCF-ResourceAgents

Loglevels:
  debug: off means log LOG_INFO and higher
  debug: on means log LOG_DEBUG and higher

The heartbeat ha_logd is not used.
In /etc/ha.d/ha.cf set: 
  logfacility	daemon
  # debug       1 only for debugging
  debug       0

Question:
I configured a lsb:heartbeat:Stateful dummy resource and did
ocf_log debug "xxx"
ocf_log info "xxx"
in a monitor op every 5s
I can see the ocf_log info in the syslog but the ocf_log debug never appears

I tried all from
debug 1
debug on
debug true
debug yes

I patched /etc/ha.d/shellfuncs to
ha_debug() {

        # SKW
        echo "`date` skw $HA_debug" >> /var/log/ocf-debug.log

        if [ "x${HA_debug}" = "x0" ] ; then
                return 0
        fi
        ....
and I found that HA_debug is always "0" whatever I put to ha.cf

What I was searching for was a way to dynamicall change the loglevel 
of ocf_log for the OCF-RAs. It seems to me that this should be 
implemented in 
.ocf-shellfuncs:ocf_log
This is in the package heartbeat-resources so I don't know if this 
is the appropriate mailing list.

For myself I found the following solution for .ocf-shellfuncs

# include another settings file
. ${OCF_ROOT}/resource.d/heartbeat/.ocf-settings
...
# prepare OCF_LOGLEVELS, called in __ocf_init_loglevel
__ocf_init_loglevel() {
    case "${OCF_LOGLEVEL}" in
      crit)	 OCF_LOGLEVELS="crit";;
      err)	 OCF_LOGLEVELS="crit err";;
      warn)	 OCF_LOGLEVELS="crit err warn";;
      info)	 OCF_LOGLEVELS="crit err warn info";;
      debug) OCF_LOGLEVELS="crit err warn info debug";;
      *)	__OCF_PRIO=`echo ${__OCF_PRIO}| tr '[a-z]' '[A-Z]'`;;
    esac
}
...
# new ocf_log with filter depending on OCF_PRI and OCF_LOGLEVEL
ocf_log() {
	# TODO: Revisit and implement internally.
	if
      [ $# -lt 2 ]
    then
      ocf_log err "Not enough arguments [$#] to ocf_log."
    fi

    __OCF_PRIO="$1"
    shift
    __OCF_MSG="$*"

    if [[ $OCF_LOGLEVELS == *$__OCF_PRIO* ]]; then
        case "${__OCF_PRIO}" in
          crit)	__OCF_PRIO="CRIT";;
          err)	__OCF_PRIO="ERROR";;
          warn)	__OCF_PRIO="WARNING";;
          info)	__OCF_PRIO="INFO";;
          debug)__OCF_PRIO="DEBUG";;
          *)	__OCF_PRIO=`echo ${__OCF_PRIO}| tr '[a-z]' '[A-Z]'`;;
        esac
	ha_log "${__OCF_PRIO}: $__OCF_MSG"
    fi 
}