Resource registration is failing - Couldnt run method tag. Error in execve

Question

Hi,
We have solaris cluster setup with two nodes and a resource group exists which contains a zfs resource which could be mounted on to one of the nodes in the cluster.
Our application is installed on one of the directory in the zfs file system. The file system and the application would be to be on one of the nodes and if somethings goes bad, it shoud failover to other node.
I used Agent Builder to create new resource type for managing our application part of sun cluster and used the ksh mode providing it start/stop/probe scripts. I have left validate script as blank.
When I try to register this new resource type, (resource time generated folder/util/start* script,
nodex /u2/appx/cluster/ResourceType_01/XYZSolarisAPPXClusterL0/util % ls -ltr
total 71
-rw-r--r-- 1 appowner appowner 13960 Jan 12 07:36 startSolarisAPPXClusterL0 <---
-rw-r--r-- 1 appowner appowner 8281 Jan 12 07:36 removeSolarisAPPXClusterL0
-rw-r--r-- 1 appowner appowner 7316 Jan 12 07:36 stopSolarisAPPXClusterL0
-rw-r--r-- 1 appowner appowner 2300 Jan 12 07:36 SolarisAPPXClusterL0_config
# Resource group creation was successful (SolarisAPPXClusterL0-harg)
# Virtual Host resource creation was successfull
# It got added to the new group created
# But it is failing when it is attempting to create resource for the new resource type - SolarisAPPXClusterL0-hars
Error:
Jan 17 14:52:01 nodex Cluster.RGM.global.rgmd: [ID 224900 daemon.notice] launching method <SolarisAPPXClusterL0_validate.ksh> for resource <SolarisAPPXClusterL0-hars>, resourcegroup <SolarisAPPXClusterL0-harg>, node <nodex>, timeout <300> seconds
Jan 17 14:52:01 nodex Cluster.RGM.fed: [ID 838032 daemon.error] SolarisAPPXClusterL0-harg.SolarisAPPXClusterL0-hars.2: Couldn't run method tag. Error in execve: No such file ordirectory.
Jan 17 14:52:01 nodex Cluster.RGM.global.rgmd: [ID 699104 daemon.error] VALIDATE failed on resource <SolarisAPPXClusterL0-hars>, resource group <SolarisAPPXClusterL0-harg>, time used: 0% of timeout <300, seconds>
The new resource type files are available on nodex and nodey. When we try to remove and start it again, the error occurred on the other nodey with similar error.
Jan 17 14:33:52 nodey Cluster.RGM.global.rgmd: [ID 224900 daemon.notice] launching method <hafoip_validate> for resource <failover-clus1>, resource group <SolarisAPPXClusterL0-harg>, node <nodey>, timeout <300> seconds
Jan 17 14:33:52 nodey Cluster.RGM.global.rgmd: [ID 515159 daemon.notice] method <hafoip_validate> completed successfully for resource <failover-clus1>, resource group <SolarisAPPXClusterL0-harg>, node <nodey>, time used: 0% of timeout <300 seconds>
Jan 17 14:33:53 nodey Cluster.RGM.global.rgmd: [ID 224900 daemon.notice] launching method <SolarisAPPXClusterL0_validate.ksh> for resource <SolarisAPPXClusterL0-hars>, resourcegroup <SolarisAPPXClusterL0-harg>, node <nodey>, timeout <300> seconds
Jan 17 14:33:53 nodey Cluster.RGM.fed: [ID 838032 daemon.error] SolarisAPPXClusterL0-harg.SolarisAPPXClusterL0-hars.2: Couldn't run method tag. Error in execve: No such file ordirectory.
Jan 17 14:33:53 nodey Cluster.RGM.global.rgmd: [ID 699104 daemon.error] VALIDATE failed on resource <SolarisAPPXClusterL0-hars>, resource group <SolarisAPPXClusterL0-harg>, time used: 0% of timeout <300, seconds>
Are there other logs than /var/adm/messages to check for errors ?
What is causing this error ? Any help is appreciated.
Thanks in advance!

Answer 1

See if adding the execute permissions to the scripts solve the problem.
In general, I just use the SUNW.gds resource type directly rather than use the agent builder. If you want to build something a little more complex, then I would recommend sub-classing the SUNW.gds agent. There are a number of Oracle supplied agents that do this, for example, the HA-LDoms, and HA-obiee agents. Have a look at how they work.
Regards,
Tim
---