Serge Dubrouski wrote:
> On 7/20/07, Andrew Beekhof <[EMAIL PROTECTED]> wrote:
>> the crm does inherit the log settings from ha.cf, however for a number
>> of reasons we default to LOCAL7 if no other value is set.
>> to disable syslog logging completely, put this in ha.cf:
>> logfacility none
> Unfortunately this isn't a true statement. I tried to set logfacility
> to local6 and then to none. Still messages in boot.log. And setting
> lofgacility to none doesn't completely disable syslog logging.
>> On 7/20/07, Serge Dubrouski <[EMAIL PROTECTED]> wrote:
>> > Just found that Heartbeat code explicitly uses LOCAL7 logfacility for
>> > system logging:
>> > cl_log_set_entity(entity);
>> > cl_log_set_facility(LOG_LOCAL7);
>> > and so on.
>> > On standard Linux distribution ( RedHat at least) LOCAL7 is dedicated
>> > for boot processes and is directed to go to boot.log. What was the
>> > reason in ignoring the logfacility value from ha.cf and explicitly
>> > using LOCAL7?
Lots of distros don't treat local7 the same. I had started using that
for my testing exactly for that reason (to keep our tests from being
confused by other system log messages). But that decision doesn't make
sense for use outside the test lab. Although I don't think I wrote any
non-test code that defaults to local7, I did set the precedent that led
to that kind of thinking.
If we are going to use some kind of default logfacility, it should
probably be daemon - which is the most generic and appropriate guess at.
We just reworked some of that code in 'dev' and it is _possible_ the
behavior has changed.
If you think 'dev' is still defaulting non-test things to LOCAL7, I
think a bugzilla would be in order. We can always do special things for
our test environment.
Alan Robertson <[EMAIL PROTECTED]>
"Openness is the foundation and preservative of friendship... Let me
claim from you at all times your undisguised opinions." - William
Home Page: http://linux-ha.org/