Hello,
I am trying to setup the snmp monitoring for our Tuxedo environments.
My comprehension of the different possibilities:
• It is possible to activate snmp management with the installation of Snmp Agent on the master of the domain.
In that case, the “tux_snmpd” process must be invoked only on the master of the Tuxedo domain with “-c” option.
• If different environments reside on the master machine, then it is necessary to invoke the process “snmp_integrator”, after what the process “tux_snmpd” must be launched (without “-c” parameter, but the option “-l” is necessary to specify the Tuxedo instance in that case).
I have installed on all machines of the 2 domains snmp_integrator and tux_snmpd (as services for Windows machines)
1. Windows environment:
An environment composed by 2 windows (W2k3) machines (win1 as master and win2, for example).
a. “single snmp agent” with the agent only running on the master
i. When I start “tux_snmpd” as agent on the master (-s –c, as parameters), the process is well loaded.
ii. When I try to walk, from the master machine, in the OID tree, I receive the following error:
E:\bea\tuxedo81\bin>snmpwalk win1 public
Mangled packet
An error occurred, Quitting
Q: What is the reason of this error ?
b. “smux configuration”, multi-domains
i. When I start “snmp_integrator” (tux_snmpd is not running), I can retrieve information as:
mgmt.mib-2.system.sysDescr.0 mgmt.mib-2.system.sysUpTime.0
But when I try to access to, by example, Tuxedo’s objects, I receive the following error:
“Error in packet.- -Reason: There is no such variable name in this MIB.”
ii. When “snmp_integrator” is running, it is impossible to launch “tux_snmpd” (-c –l simpapp or -c –s –l simpapp, as parameters)
Q: What is the reason of that comportment ?
2. Solaris and window cohabitation
A Tuxedo domain composed by:
• Solaris1, as master of the domain (Solaris) which is admin of different Tuxedo domains
• Win1, as application machine of test environment (W2k or W2k3)
• Win2, as application machine of acceptation environment (W2k or W2k3)
1. When launched of “snmp_integrator”, the process go in “running” state and take more than 80% of CPU of the Solaris machine, and carries out to nonviable system.
Our main concern is in the “Solaris and window cohabitation”.
Q: Could you have had also this comportment and do you know how resolve this case?
I am trying to setup the snmp monitoring for our Tuxedo environments.
My comprehension of the different possibilities:
• It is possible to activate snmp management with the installation of Snmp Agent on the master of the domain.
In that case, the “tux_snmpd” process must be invoked only on the master of the Tuxedo domain with “-c” option.
• If different environments reside on the master machine, then it is necessary to invoke the process “snmp_integrator”, after what the process “tux_snmpd” must be launched (without “-c” parameter, but the option “-l” is necessary to specify the Tuxedo instance in that case).
I have installed on all machines of the 2 domains snmp_integrator and tux_snmpd (as services for Windows machines)
1. Windows environment:
An environment composed by 2 windows (W2k3) machines (win1 as master and win2, for example).
a. “single snmp agent” with the agent only running on the master
i. When I start “tux_snmpd” as agent on the master (-s –c, as parameters), the process is well loaded.
ii. When I try to walk, from the master machine, in the OID tree, I receive the following error:
E:\bea\tuxedo81\bin>snmpwalk win1 public
Mangled packet
An error occurred, Quitting
Q: What is the reason of this error ?
b. “smux configuration”, multi-domains
i. When I start “snmp_integrator” (tux_snmpd is not running), I can retrieve information as:
mgmt.mib-2.system.sysDescr.0 mgmt.mib-2.system.sysUpTime.0
But when I try to access to, by example, Tuxedo’s objects, I receive the following error:
“Error in packet.- -Reason: There is no such variable name in this MIB.”
ii. When “snmp_integrator” is running, it is impossible to launch “tux_snmpd” (-c –l simpapp or -c –s –l simpapp, as parameters)
Q: What is the reason of that comportment ?
2. Solaris and window cohabitation
A Tuxedo domain composed by:
• Solaris1, as master of the domain (Solaris) which is admin of different Tuxedo domains
• Win1, as application machine of test environment (W2k or W2k3)
• Win2, as application machine of acceptation environment (W2k or W2k3)
1. When launched of “snmp_integrator”, the process go in “running” state and take more than 80% of CPU of the Solaris machine, and carries out to nonviable system.
Our main concern is in the “Solaris and window cohabitation”.
Q: Could you have had also this comportment and do you know how resolve this case?