SUNW.gds is trying to validate start up command on both nodes of cluster

Question

I have two nodes configured to do a failover rg which contains SMC 4.0 server.
I have the ZFS pool set up as a HASP resource
I also have a Logical Hostname resource configured
I have installed SMC 4.0 in /opt/smc (which is a zfs mountpoint)
I run the following command --
clresource create  -g scha_smc_rg -t SUNW.gds:6 \
-p Scalable=false  -p Start_command="/opt/smc/SUNWsymon/sbin/es-start -cA" \
-p Stop_command="/opt/smc/SUNWsymon/sbin/es-stop -A" \
-p Port_list="8080/tcp" -p Failover_enabled=TRUE \
-p Child_mon_level=0 -p network_resources_used=smclhn-rs scha_smcIt validates on the node on which I have the rg running at the moment.
It refuses to validate on the other node (standby node), which is very obvious since the Filesystem and app can be running only on one node at a time.
Apr  3 11:44:41 ccs-01 SC[,SUNW.gds:6,scha_smc_rg,scha_smc_rs,gds_validate]: [ID 586689 daemon.error] Cannot access the start command </opt/smc/SUNWsymon/sbin/es-start> : <No such file or directory>Any idea why this is happening? I'm trying to set this app up as a Failover resource not a scalable one.

Answer 1

I spent many months c2002 trying to get SunMC 3.x (??) to work in a fail-over environment. This proved to be fraught with problems. The best solution is to put SunMC into a fail-over container instead and start it via SMF.
The reason this is the easiest option is that the IP address and hostname of the container doesn't change on fail-over, whereas the hostname does if it is run in the global zone. The hostname is used inside SunMC (IIRC) to bind various 'bits and piece' (don't remember the detail) together.
Regards,
Tim
---