bounce_logger – Momentum-Style Bounce Logging
The ec_logger and event_hydrant modules provide full disposition status of every message that goes through Momentum. However, it is often convenient to view status of only bounced messages. The bounce_logger module writes only inband (or protocol-time) and out-of-band bounces to a single log file.
For details about the bouncelog file format, see “bouncelog”.
To configure the bounce_logger module, load it in the ecelerity.conf file. The following is the default configuration:
bounce_logger "bounce_logger"
{
bouncelog = "/var/log/ecelerity/bouncelog.ec"
bouncelog_mode = 0644
heartbeat = 60
}
This configuration instructs Momentum to write a bouncelog file to /var/log/ecelerity/bouncelog.ec with the file permissions 0644.
The heartbeat option sets the interval at which "heartbeat" entries will write to the log. Default value is 60. The heartbeat entries in the log are essential for proper operation of the real time stats pieces, and changing the default value of heartbeat is not recommended unless advised to do so by Message Systems support.
Note
If you find that you have rejectlog entries with the reason "550 5.7.1 [internal] discarded by policy", and you are having difficulty in correlating those rejections with your policy rules, it could be because your bounce processing configuration is set to blackhole any detected bounces. If this is the case, you should expect to see a corresponding entry in your bounce log, except in the case where the internal bounce classifier has determined that the incoming message was classified as BC_SUBSCRIBE, a subscription request. The bounce logger does not log this type of incoming mail because it does not relate to a specific outgoing message originating from your site.
In the configuration, you can also specify which events to log or not to log. Events will be logged based on the following lines in your configuration file:
log_inband_bounces = <on|off: default on>
log_outofband_bounces = <on|off: default on>
log_transient_failures = <on|off: default on>
The log_transient_failures option creates a log entry of type ‘T’ in the bounce log along with the actual bounce log line. Apart from the type identifier, the log entry for a transient failure is the same as other bounce log entries. Note that bounce log entries differ in format from ec_logger entries. The default value for this option is on.
The bounce_logger relies on additional global configuration options: bounce_domains, bounce_behavior, bounce_suppress_list and bounce_pattern.
The default log file created by this logger is rotated by the utility script ec_rotate. For more information, see ec_rotate.
When the bounce_logger is defined in the ecelerity-cluster.conf file, it configures the files used to create aggregated text logs of node events.
The default ecelerity-cluster.conf file defines the following bounce_logger:
bounce_logger "bounce_logger_cluster" { bouncelog = "cluster:///var/log/ecelerity/bouncelog.cluster=>master" }
For a detailed explanation of setting up cluster-wide consolidated logging, see “Configuration of Aggregated Cluster Node Logging”.
The bounce_logger module can also be used to configure log events that occur on the cluster manager. Configuration is similar to “Configuration” with the exception that the bounce_logger module is loaded in the eccluster.conf file.
The bounce_logger module allows for a limited set of online commands via ec_console:
- bounce_logger:*`bounce_logger_rt`* reopen logs
-
If you move a log, use the ec_console command reopen logs to close and reopen all the log files.