A very short section which just doesn't seem to belong anywhere else.
Epos supports error logging via syslog (on systems where syslogd
is running). At the moment, only (the most of) TTSCP completion codes
together with the associated messages can be logged; this includes
all abnormal server termination states (except for a plain old fatal
signal). If syslog should be available (the syslog.h
header
file is present), but isn't, the error messages are logged directly
to the system console. If the syslog facility is being missed at
compilation time, Epos opens a file called epos.err
in the
current directory whatever it may be and writes some debugging
information there instead.
The syslog message levels are set according to the
syslog message levels table. In all cases, Epos messages
are logged with the daemon
facility.
level | used for |
debug | trivial success TTSCP codes |
info | silly errors and other messages |
notice | shutting down as requested |
warning | security or network error |
err | internal error |
crit | abnormal termination |
See also the <tt>authpriv</tt> option which may change this behavior for possibly security-related errors, but not network errors.
Any debugging output together with error messages etc. is also often logged
to /var/log/epos
as specified with the
<tt>daemon_log</tt> option.
The agent profile which shows how much time does Epos spend with individual agents (which gives an insight into which processing phases are time consuming) can be turned on with the <tt>profile</tt> option.