Event Categories

This section describes the events and the different event categories the Oracle Communications Session Border Controller can generate.

About Events

Events are the circumstances that generate one or more of the following:

  • alarm
  • entry in a log file
  • SNMP trap

The following table lists the three categories used to define these events.

Event Category Description
Informational Represents non-critical conditions. For example, a configuration element has changed.
Warning Indicates pending failures or unexpected events. For example, you typed the wrong password at the console three consecutive times.
Error Indicates that a serious condition has occurred. For example, an internal temperature reading exceeds the recommendation.

These broad categories generally consist of the facility that generated them, along with an indication of the severity of the message. This information helps filter the more important and time-sensitive notifications from the informative messages.

Types of Events

The Oracle Communications Session Border Controller can generate the following types of events.

  • process log events
  • system log events
  • protocol trace elements

Process Log Events

Events are logged to the process log flow from tasks and are specific to a single process running on the Oracle Communications Session Border Controller. By default they are placed into individual files associated with each process with the following name format:

log.<taskname>

Note:

Process logs serve as a debugging tool. When set to debug level, the quantity of events generated can become overwhelming for the Oracle Communications Session Border Controller. It should only be used by Oracle personnel, or with their assistance. It is not recommended for use on production systems.

When you configure the system, you set the default system-wide process log level and each task logs according to this setting. You can override this log level for specific tasks when configuring other elements. For example, when you configure the media manager you can set the ALGD and MBCD log levels to different severity levels.

System Log Events

System log events are a subset of the collection of all process log events. Every software process writes messages to a file called acmelog, if the severity of the event meets or exceeds the configured log level threshold. There is one system log for the whole Oracle Communications Session Border Controller (filename: acmelog).

System log events are also referred to as acmelog events and are analogous to a traditional syslog event. The acmelog file is typically viewed as an aggregation of notable alarms and errors from all software processes.

The Oracle Communications Session Border Controller supports logging using SYSLOG, which is an industry-standard protocol that lets a device send event notification messages across IP networks to event message collectors - also known as syslog servers. Messages are usually sent using UDP port 514.

The Oracle Communications Session Border Controller can send information to a remote SYSLOG server. You configure the server and globally set the severity level at which the Oracle Communications Session Border Controller logs events when you configure the system.

Protocol Trace Events

Protocol trace events are the events associated with a protocol transaction. They are enabled on a per-process basis using the notify command, resulting in transactional events being placed into transaction logs, such as sipmsg.log.

These events are helpful for troubleshooting sessions, but they are also the highest volume events the Oracle Communications Session Border Controller produces and can only be enabled for short times.

Event Granularity

You can set the reporting level for events placed into the logs by using the following methods:

  • Setting the system-wide severity level (at or above which events are logged) by configuring the system’s process log level. This setting is persistent across boots.

    You set the system-wide severity level by configuring the log severity level threshold when performing the system configuration.

  • Configuring individual parameters for different elements that control specific process logs. For example, you can configure the mbcd log level for the media manager. These settings are persistent across boots.

    For example, to configure the process log level for monitoring all H.323 activity on the Oracle Communications Session Border Controller, you configure the log level to INFO when configuring H.323 signaling.

  • Using ACLI log-level command to dynamically specify the log level for a specific task (or all tasks using the keyword all). You can specify finer granularity by including specific subtypes within the process. These settings are not persistent across boots.
  • Using the ACLI or Acme Control Protocol (ACP) notify command. For example, notify mbcd debug. Such settings are not persistent across boots.

Event Severity

There are eight severity levels ranging from lowest severity, Debug, to the highest, Emergency.

syslog Numerical Code syslog Severity Oracle Log Enumeration
0 Emergency (system is unusable) EMERGENCY (0)
1 Alert (action must be taken immediately) CRITICAL (1)
2 Critical (critical conditions) MAJOR (2)
3 Error (error conditions) MINOR (3)
4 Warning (warning conditions) WARNING (4)
5 Notice (normal but significant condition) NOTICE (5)
6 Informational (informational messages) INFO (6)
7 Debug (debug level messages) TRACE (7)

DEBUG (8)

DETAIL (9)

SNMP Traps

The Oracle Communications Session Border Controller supports several standard SNMP traps (cold start, link up/down) and proprietary traps used to notify SNMP managers of specific events:

  • apSysMgmtGroupTrap – used for different events. The trap must be parsed by a management tool to extract the specific event details.
  • specific uniquely identified traps – used for specific Oracle Communications Session Border Controller events. These traps correspond exactly to the events that show up in acmelog.

    The unique traps are only generated if the system-config, enable-snmp-monitor-traps ACLI parameter is enabled:

  • apSysLogGeneratedTrap – used as a catch-all for system log (syslog) events.

See the MIB Reference Guide for more details about traps.

Alarms

The most serious events noted by the Oracle Communications Session Border Controller are categorized as alarms. They appear in the alarm table, which is displayed in the ACLI using the command display-alarms. The ACLI also supports clearing alarms displayed in that table. Alarms are not sent off-box explicitly, however, at least one of the following mechanisms is usually triggered when an alarm occurs:

  • A dry contact port on the back of the chassis that may be used to control a remote alarm panel.
  • An SNMP trap may be generated
  • A syslog event may be generated

See the MIB Reference Guide for details about alarms.

Process Log Events

Process log events can be sent to a log server by configuring the system to include the destination server’s IP address and port number. For example, using the ACLI you configure the following system parameters:

  • process-log-server
  • process-log-port

    The process log port can be any port from 1025 to 65535. It is most commonly configured as port 2500.

The Oracle Communications Session Border Controller stops logging events to RAM memory and instead sends them to the configured remote server over UDP. Because of the added overhead of sending log messages using UDP datagrams versus writing to the RAM drive, message content decreases – even at the same configured log levels.

System Log Events

System log events can be sent to one or more syslog servers using the traditional UNIX syslog mechanism as described in RFC 3164. Users can configure one or more syslog servers to which the Oracle Communications Session Border Controller will send generated syslog events by setting the following syslog parameters in the system configuration:

  • address
  • port
  • facility

If the port is left empty, the default value is UDP port 514 (the well-known syslog port).

Traps

Traps are defined to be sent to a SNMP Manager using the following configuration parameters:
  • system-config, trap-receiver, ip-address
  • system-config, trap-receiver, filter-level
  • system-config, trap-receiver, community-name

Alarms

Alarms can be sent off the box using the dry contact port in the rear of the chassis.