ajb007
July 14, 2020, 8:12pm
22
I have own solution of this bug.
Syslog-ng have options
keep-timestamp which tell (when NO) set(write) actual time(real) of message to the log file.
I have create file /etc/syslog-ng.d/correct-time.conf
with content
options {
keep-timestamp(no);
};
And for me is this best solution.
3 Likes
Do you have a (documentation) link about that supposedly TZ logging?
According to RFC 5424 - The Syslog Protocol and other sections of that RFQ there is no mandatory TZ aware logging by the source/originator.
This document describes the syslog protocol, which is used to convey event notification messages. This protocol utilizes a layered architecture, which allows the use of any number of transport protocols for transmission of syslog messages. It also...
It is recommended that the value of “0” be the default for the “tzKnown”
RFC5424 describes internet protocol
syslog on all UN*X systems I know work with local timezone
dnsmasq, syslog-ng, sentinel and possibly others work with local time, so there’s at least inconsistency here.
yes, it can be fixed at syslog level using:
would work, at least until the problem itself is fixed.
is nice, but the log above shows that odhcpd doesn’t use busybox nor procd logger
It describes the protocol that is utilised by syslog() call
This document describes the syslog protocol, which is used to convey event notification messages.
fantomas:
dnsmasq, syslog-ng, sentinel and possibly others work with local time, so there’s at least inconsistency here.
Which may very well because one userland implemented tzKnown="1"
whilst other userland did not bother and defaults to tzKnown="0"
instead or is coded to emit in UTC in any case.