Hi all,
I am about to release 3.21.4, BUT it contains one big change and I would
appreciate if some of you could give it a try before I officially
release it. In 3.21.4, I have removed the compile-time hardcoded maximum
message size. So far, a message, by default, could be 2K maximum. This
could be changed by modifying the MAXLINE compile time default (and
recompiling/install). This size limit stems back to plain old BSD
syslogd, RFC 3164 (where it is 1k) and the upcoming syslog RFC series
(where it is variable, but 2k is suggested as a minimum). There have
been a lot of questions about larger sizes. So I have now streamlined
the process and created a simple config directive:
$MaxMessageSize 4k
Where 4k is of your choice. You may also specify any non-k value, like
7341. I do not check for a sensible maximum, so you may also specify 4p
(4 petabyte), but I hardly believe that will get you a working instance
;)
I did not blindly change the message size - I tried to keep performance
as good as with the previous approach (dyn memory alloc takes more
time!). So there is chance for failure all over the code. In my test (up
to now), everything works, but I would really appreciate some feedback
from practice.
Please let me know if you experience any issues with that release.
It is available at
http://download.rsyslog.com/rsyslog/rsyslog-3.21.4-Test1.tar.gz
This version also contains a number of smaller fixes, but not yet the
new doc building system (this is next on my agenda).
Your feedback is appreciated.
Rainer
I am about to release 3.21.4, BUT it contains one big change and I would
appreciate if some of you could give it a try before I officially
release it. In 3.21.4, I have removed the compile-time hardcoded maximum
message size. So far, a message, by default, could be 2K maximum. This
could be changed by modifying the MAXLINE compile time default (and
recompiling/install). This size limit stems back to plain old BSD
syslogd, RFC 3164 (where it is 1k) and the upcoming syslog RFC series
(where it is variable, but 2k is suggested as a minimum). There have
been a lot of questions about larger sizes. So I have now streamlined
the process and created a simple config directive:
$MaxMessageSize 4k
Where 4k is of your choice. You may also specify any non-k value, like
7341. I do not check for a sensible maximum, so you may also specify 4p
(4 petabyte), but I hardly believe that will get you a working instance
;)
I did not blindly change the message size - I tried to keep performance
as good as with the previous approach (dyn memory alloc takes more
time!). So there is chance for failure all over the code. In my test (up
to now), everything works, but I would really appreciate some feedback
from practice.
Please let me know if you experience any issues with that release.
It is available at
http://download.rsyslog.com/rsyslog/rsyslog-3.21.4-Test1.tar.gz
This version also contains a number of smaller fixes, but not yet the
new doc building system (this is next on my agenda).
Your feedback is appreciated.
Rainer