Serviceguard service problem

[jpcast_real]

Hello,

I have added a new Service to my Package and the startup of this new service is really strange . The new service start just after de filesystems mount step and this makes the PKG goes down because the "customer defined" section has not still been executed. With the other two services the behauviour is normal, any idea?????

SERVICE_NAME oracle-monitor
SERVICE_FAIL_FAST_ENABLED NO
SERVICE_HALT_TIMEOUT 5

SERVICE_NAME scsupvdb-monitor
SERVICE_FAIL_FAST_ENABLED NO
SERVICE_HALT_TIMEOUT 5

SERVICE_NAME hidb-monitor
SERVICE_FAIL_FAST_ENABLED NO
SERVICE_HALT_TIMEOUT 5




SERVICE_NAME[0]=oracle-monitor
SERVICE_CMD[0]="/etc/cmcluster/scripts/monitor-oracle"
SERVICE_RESTART[0]="" # Will not restart the service.

SERVICE_NAME[1]=scsupvdb-monitor
SERVICE_CMD[1]="/etc/cmcluster/scripts/monitor-scsupvdb"
SERVICE_RESTART[1]="-R"

SERVICE_NAME[2]=hidb-monitor
SERVICE_CMD[2]="/etc/cmcluster/scripts/monitor_hi_pkgdb"
SERVICE_RESTART[2]="-R" # Will not restart the service.

[Mark Nieuwboer]

Hi,

Can you give the log file.
This is very little to say something about it. Normaly it's something it can exucute the monitor.

grtz. Mark

[jpcast_real]

Hello Mark,

The log file is really big and it is difficult to understand if you are not familiarized with our application . In this moment I am making a test just with one service ,the hidb-monitor service wiht the instanace 0 , and works perfectly. I have also tested with 2 services , instance 0 and 1:

Athos:/etc/cmcluster/pkgdb> more *log | grep "Jan 20 10:08"
########### Node "Athos": Starting package at Fri Jan 20 10:08:46 MET 2006 ###########
Jan 20 10:08:46 - "Athos": Activating volume group vgora with exclusive option.
Jan 20 10:08:47 - "Athos": Activating volume group vgredo with exclusive option.
Jan 20 10:08:47 - "Athos": Activating volume group vgpkgdb with exclusive option.
Jan 20 10:08:47 - Node "Athos": Checking filesystems:
Jan 20 10:08:48 - Node "Athos": Mounting /dev/vgpkgdb/FLIP at /mad503
Jan 20 10:08:48 - Node "Athos": Mounting /dev/vgpkgdb/FLOP at /mad503-alt
Jan 20 10:08:48 - Node "Athos": Mounting /dev/vgora/data at /mad503/madora
Jan 20 10:08:48 - Node "Athos": Mounting /dev/vgredo/redo at /mad503/redo
Jan 20 10:08:49 - Node "Athos": Mounting /dev/vgpkgdb/trace at /mad503/output
Jan 20 10:08:50 - Node "Athos": Adding IP address 174.1.51.55 to subnet 174.1.51.0
SQL*Plus: Release 9.2.0.5.0 - Production on Fri Jan 20 10:08:50 2006
Jan 20 10:08:57 - Node "Athos": Starting service oracle-monitor using
Jan 20 10:08:57 - Node "Athos": Starting service hidb-monitor using
########### Node "Athos": Package start completed at Fri Jan 20 10:08:57 MET 2006 ###########

[Mark Nieuwboer]

Hi,

what does the syslog say.
And what kind of error messages did you get in the logfiles.
Futher more what is the new Service i think monitor_hi_pkgdb.

good luck whit yur test.

grtz. Mark

[jpcast_real]

the new service is hidb-monitor , I have changed the name of the executable file :

SERVICE_NAME[2]=hidb-monitor
SERVICE_CMD[2]="/etc/cmcluster/scripts/monitor-hidb"
SERVICE_RESTART[2]="-R"

Now things are working , it seems that the new executable name "monitor-hidb" it works...

Do you know if there is any kind of naming limitation for the PKGs ??

[Mark Nieuwboer]

Not that i know of.
I had longer names for services.
Maybe something to do with the underscores.

grtz. Mark

[Victor Fridyev]

Hi,

It seems to me, you have to delay start up of your package until the DB package will be in status "running".

HTH