Not sure if you can use this, but I had an escalation where a legacy
script was taking too long during boot, we suspected the script
was in rc2.d, so to timestamp each script , I had the customer
change the existing $action case statement within /etc/rc2 as
described below::

First make a backup of the file /etc/rc2, then change the start
section of the $action case statement from this:

        start)
                for f in /etc/rc2.d/S*; do
                        if [ -s $f ]; then
                                case $f in
                                        *.sh)   /lib/svc/bin/lsvcrun -s $f \
                                                        start ;;
                                        *)      /lib/svc/bin/lsvcrun $f start ;;
                                esac
                        fi
                done
                ;;

        *)
                echo "Invalid argument"
                exit 1
esac
exit 0

-------------
TO THIS:
-------------

        start)
                for f in /etc/rc2.d/S*; do
                        if [ -s $f ]; then
                                case $f in
                                *.sh)   echo $f >> /outfile ; /bin/date >> 
/outfile 2>&1 ; /lib/svc/bin/lsvcrun -s $f st
art ;;
                                *) echo $f >> /outfile ; /bin/date >> /outfile 
2>&1 ; /lib/svc/bin/lsvcrun $f start ;;
                                esac
                        fi
                done
                ;;

        *)
                echo "Invalid argument"
                exit 1
esac
exit 0


After reboot the resultant /outfile file 
will resemble something like this:

/etc/rc2.d/S10lu
Wed Nov  1 18:44:41 EST 2006
/etc/rc2.d/S20sysetup
Wed Nov  1 18:44:45 EST 2006
/etc/rc2.d/S40llc2
Wed Nov  1 18:44:46 EST 2006
/etc/rc2.d/S42ncakmod
Wed Nov  1 18:44:46 EST 2006
/etc/rc2.d/S47pppd
Wed Nov  1 18:44:46 EST 2006
/etc/rc2.d/S70sckm
Wed Nov  1 18:44:47 EST 2006
/etc/rc2.d/S70uucp
Wed Nov  1 18:44:47 EST 2006
/etc/rc2.d/S72autoinstall
Wed Nov  1 18:44:47 EST 2006
/etc/rc2.d/S72directory
Wed Nov  1 18:44:47 EST 2006
/etc/rc2.d/S73cachefs.daemon
Wed Nov  1 18:44:47 EST 2006
/etc/rc2.d/S81dodatadm.udaplt
Wed Nov  1 18:44:47 EST 2006
/etc/rc2.d/S89PRESERVE
Wed Nov  1 18:44:48 EST 2006
/etc/rc2.d/S89bdconfig
Wed Nov  1 18:44:48 EST 2006
/etc/rc2.d/S90wbem
Wed Nov  1 18:44:48 EST 2006
/etc/rc2.d/S90webconsole
Wed Nov  1 18:44:49 EST 2006
/etc/rc2.d/S91afbinit
Wed Nov  1 18:44:58 EST 2006
/etc/rc2.d/S91gfbinit
Wed Nov  1 18:44:59 EST 2006
/etc/rc2.d/S91ifbinit
Wed Nov  1 18:44:59 EST 2006
/etc/rc2.d/S91jfbinit
Wed Nov  1 18:45:00 EST 2006
/etc/rc2.d/S91zuluinit
Wed Nov  1 18:45:00 EST 2006
/etc/rc2.d/S94ncalogd
Wed Nov  1 18:45:00 EST 2006
/etc/rc2.d/S95IIim
Wed Nov  1 18:45:00 EST 2006
/etc/rc2.d/S97cc-transport
Wed Nov  1 18:45:01 EST 2006
/etc/rc2.d/S98cc-ccragent
Wed Nov  1 18:45:02 EST 2006
/etc/rc2.d/S98cc-fwagent
Wed Nov  1 18:45:02 EST 2006
/etc/rc2.d/S98deallocate
Wed Nov  1 18:45:02 EST 2006
/etc/rc2.d/S99cc-invagent
Wed Nov  1 18:45:03 EST 2006
/etc/rc2.d/S99dtlogin
Wed Nov  1 18:45:03 EST 2006
/etc/rc2.d/S99vnc
Wed Nov  1 18:45:04 EST 2006


After reboot, the script that is taking too long time should be evident.



-ED




Renaud Manus wrote:

> Are you saying that you would like the output of your own rc scripts 
> to be
> displayed on the console rather than in the log files?
> If it's the case then the SMF verbose logging option is not going to 
> help.
>
> What you should do is modify your rc scripts so they explicitly output 
> to the
> console.
>
> eg. add the following at the beginning of your scripts:
> [ -x /lib/svc/bin/svc.startd ] && exec 1> /dev/sysmsg 2>&1
>
> Regards,
> Renaud
>
> Sean Liu wrote:
>
>> When doing jumpstarts, to install 3rd party softwares, a few RC 
>> scripts were created and run after the system reboot. A few reboots 
>> would then be performed at different stages ( adding packages, adding 
>> patches, encap root, mirror root etc.etc. )
>>
>> For solaris 9 and below, you'll see the rc scripts run and know 
>> instantly when something goes wrong.
>> However with Solaris 10 those messages don't show up on console and 
>> you have to check the logs to find out where you are, which is an 
>> inconvinience to state the least.
>>
>> Is there a way to turn on SMF verbose output at the end of 
>> installation but BEFORE the system reboots? So at the end of all 
>> reboots the script can simply turn off the verbose output.
>>
>> Thanks,
>>
>> Sean
>>  
>>  
>> This message posted from opensolaris.org
>> _______________________________________________
>> smf-discuss mailing list
>> smf-discuss at opensolaris.org
>
>
> _______________________________________________
> smf-discuss mailing list
> smf-discuss at opensolaris.org



Reply via email to