Re: Custom Health Checks, Logging error, & Integrating w/Keepalived

2009-12-23 Thread Willy Tarreau
On Tue, Dec 22, 2009 at 01:25:24PM -0800, Paul Hirose wrote:
(...)
> Next, I think this is related to what Willy warned about in a previous
> thread about the default syslogd being slow.  I fired off a giant
> batch of requests at my haproxy, and there were a lot of error/warning
> msgs on my window about log and no resources or something.  Since the
> msgs didn't actually get logged but simply printed on the screen, I
> don't recall what that was.

On many systems, the "emerg" messages are sent on all consoles but are
also logged. You can limit the level of syslog messages in haproxy's
configuration (last parameter of the "log" statement). For instance,
if you put "alert" there, all crit and emerg messages will be sent as
"alert".

> incidentally, why is it "dontlognull" and not "dontlog-null" just to
> match the "dontlog-normal" syntax?

Because it was chosen earlier and at that time I found it easy to have
short (unreadable) names, which only became an issue once there were
many of them. It's the same for "srvtimeout", "option forwardfor",
"option httpchk" and such horrible things that people can't always
spell.

Regards,
Willy




Re: Custom Health Checks, Logging error, & Integrating w/Keepalived

2009-12-23 Thread Willy Tarreau
Hey Paul,

it's hard to reply to your mails, you send 10 questions at once. If it
takes an hour to reply to each mail, there's very little chance anyone
will find that time. Please try to split them by precise subjets and
expect that some of them will not get any reply :-)

I'll try to address some of your questions soon, maybe this week-end.

Regards,
Willy