Download This Guide
Auto-Restart Files
When enabled, the auto-restart module (radiusd script), acts as a watchdog daemon monitoring the status of the SBR Carrier executable, and restarting it as needed. Auto-restart is disabled by default. Regardless of whether auto-restart is enabled, the SBR Carrier software is always started and stopped using this command:
/opt/JNPRsbr/radius/sbrd start radius
/opt/JNPRsbr/radius/sbrd stop radius
Auto-restart is enabled by configuring the WATCHDOG* parameters in the sbrd.conf file, and fine-tuned by configuring the radiusd.conf file. See sbrd.conf File, and radiusd.conf File.
Perl must be installed on the Steel-Belted Radius Carrier server if you want to use the automatic restart module.
![]() | Note: If Perl version 5 is not installed, the radiusd script will not run, even if enabled by configuration, and SBR Carrier will operate without the auto-restart module running. |
Perl SNMP Support
You can configure the auto-restart module to send SNMP traps to record auto-restart events. Perl SNMP support resides in the Perl SNMP_Session module, which provides access to remote SNMP agents.
Perl SNMP support allows Steel-Belted Radius Carrier to send SNMP traps to a variety of SNMP agents, including the Sun Management Center, which is distributed with some Sun hardware platforms. Sun Management Center is not required to run radiusd.
Perl System Log Support
The optional perl package syslog.ph is used to log the watchdog daemon status. You can configure auto-restart to send system log messages to record auto-restart events. To use system log reporting, you can use the h2ph utility to create a syslog.ph file. This example assumes site_perl/5.005 is in @INC:
su - root cd /user/include/sys /usr/perl15/bin/h2ph -d /usr/perl15/site_perl/5.005 syslog.h
If you do not want to use system log, use the -d or --logfile options for the radiusd command to open a regular log file (radiusd.log).
sbrd.conf File
To enable the auto-restart module, you must edit the sbrd.conf file using the following procedure:
![]() | Note: The following procedure is dependent upon you enabling the auto startup scripts when you install SBR Carrier. |
- If SBR Carrier is already running, become superuser and
type this command to stop the server:
/opt/JNPRsbr/radius/sbrd stop radius
- Edit sbrd.conf to set WATCHDOGENABLE=1 (or 0 to disable)
- Type this command to restart the server:
/opt/JNPRsbr/radius/sbrd start radius
Note: If you are running the optional SSR module, the command to start/stop only the SSR cluster is /opt/JNPRsbr/radius/sbrd start/stop ssr. This controls SSR without starting or stopping the SBR.
radiusd.conf File
The default radiusd.conf settings cause the auto-restart feature to work this way:
If the radius server executable fails to respond to status polling from radiusd within 17 seconds, radiusd attempts to stop radius using SIGTERM (a polite shutdown). If radius does not shut down within 60 seconds, SIGKILL (a hard kill) is used to stop it. After shutdown by either method, radiusd starts a new radius child process. If this radius child does not respond to status polling within 60 seconds of startup, it is presumed dead; a misconfiguration of the server is assumed; and radiusd terminates with a critical error.
While the auto-restart module is enabled, all informational, debugging, warning, error, and critical messages from radiusd are recorded here:
- Syslog—Messages are written to the syslog system logging facility.
- Log file—If syslog is not available, messages are written to the
server log file specified using the --logfile option in the sbrd.conf WATCHDOGOPTS
parameter.
Note: If Perl is not installed in the
/usr/local/bin/
directory, this error message occurs when you start the SBR Carrier server:
./S90sbrd: /RadiusHome/radiusd: not found
To fix this error, edit the first line of the radiusd file in the RADIUS directory so that the directory structure points to the correct Perl interpreter executable:
#!/usr/local/bin/perl
radiusd.conf Configuration File
The radiusd.conf configuration file (Table 38) provides settings for the radiusd auto-restart module.
Table 38: radiusd.conf Syntax
radiusd.conf Parameter | Function |
---|---|
WatchdogIntervalPing | Number of seconds the automatic-restart module waits between sending status inquiries. Default value is 5 seconds. |
WatchdogIntervalMaxPong | Number of seconds the automatic-restart module waits for a reply before issuing a SIGTERM (shutdown) message. Default value is 17 seconds. |
WatchdogIntervalMaxStartup | Number of seconds during which the server is expected to be able to start up. Default value is 60 seconds. |
WatchdogIntervalMaxShutdown | Number of seconds during which the server is expected to be able to shut down. Default value is 60 seconds. |
SnmpManager = hostname community port version | Identifier for an SNMP management station you want to receive traps from the automatic-restart module. You can specify more than one SNMP management station. For each SNMP management station, enter:
|
SnmpInterface | Identifies the IP network interface to be used to generate SNMP trap messages. You can specify interfaces by name or by IP address. If you enter any, the first IPv4 interface the automatic-restart module finds is used. If you leave this parameter blank, generation of SNMP trap messages is disabled. |
SnmpCommandTrap | Specifies how SNMP trap messages are forwarded:
|
SnmpCommandUptime | Specifies how the automatic-restart module determines elapsed time for timestamps in trap messages. You can specify the pathname and filename for a module or executable
whose syntax matches the SMC uclock utility.
For example: If you leave the parameter blank, the automatic restart module calculates elapsed time relative to its own start time. Default value is blank. |
SnmpEnterprise | Specifies the OID prefix for enterprise-specific trap messages, which is used to select the appropriate MIB for decoding traps. Default value is 1.3.6.1.4.1.1411.1.1. If you leave the parameter blank, SNMP trap messages are not generated. |
SnmpGenericTrapType= 6 | Specifies the enterprise-specific trap type, which must be 6 according to the SNMPv1 standard. Do not change this value without a specific reason. |
SnmpTrapWatchdogStarted | Specifies the trap type for messages indicating that the automatic-restart module is started. Default value is 113. Enter 0 to disable this type of trap. |
SnmpTrapWatchdogStopped | Specifies the trap type for messages indicating that the automatic-restart module is stopped. Default value is 114. Enter 0 to disable this type of trap. |
SnmpTrapWatchdogRadius | Specifies the trap type for messages indicating that the RADIUS server is restarted. Default value is 115. Enter 0 to disable this type of trap. |
SnmpTrapWatchdogRadiusTerm | Specifies the trap type for messages indicating that the RADIUS server is not responding and that the automatic-restart module has sent the SIGTERM signal. Default value is 5028. Enter 0 to disable this type of trap. |
SnmpTrapWatchdogRadiusKill | Specifies the trap type for messages indicating that the RADIUS server is not responding and that the automatic-restart module has sent the KILL signal. Default value is 5029. Enter 0 to disable this type of trap. |
SnmpTrapWatchdogAborted | Specifies the trap type for messages indicating that the RADIUS server is not responding and that the automatic-restart module has given up and aborted. Default value is 10051. Enter 0 to disable this type of trap. |
SnmpTrapWatchdogFailedInit | Specifies the trap type for messages indicating that the automatic-restart module failed to start, which may indicate a misconfiguration issue. Default value is 10052. Enter 0 to disable this type of trap. |