Bug 28634

Comment 1 dwmalone 2001-07-02 16:16:42 UTC

Responsible Changed
From-To: freebsd-bugs->dwmalone

I'll have a look at this as I imported the hostname stuff.

On Mon, Jul 02, 2001 at 08:00:23AM -0700, Andy said:
> 
> When using the "log from hostname" function in syslog.conf like:-
> 
> #+hostname
> local7.=info  <tab> <tab>  /var/log/file
> #+*
> 
> fails to log to file correctly, in fact, logs nothing.

I've been having the exact same problem :

*.*                                             /var/log/all.log
+212.250.77.217
*.*                                             /var/log/netpilot

logs all messages from 212.250.77.217 to /var/log/all.log instead of
/var/log/netpilot.

Ceri

-- 
Your local RFC Nazi

It's simple to see what's going on without
using another host to syslog from/to. Just
create a syslog.conf file with the requited
spec. Then when running syslogd use -d option
and see the line cfline where it gets parsed.
You'll see it includes the leading "+" char.
My patch just skipped by one char to drop it.

I compiled and ran this and it worked straight
off.

Regards
Andy

Comment 4 dwmalone 2001-07-02 19:22:32 UTC

State Changed
From-To: open->analyzed

I think the problem here is that giving the full hostname doesn't work 
if the remote machine is in the same domain as the machine running syslogd. 
I'll try to fix this and improve the explaination in the man page.

Comment 5 ru 2001-07-03 09:42:24 UTC

On Mon, Jul 02, 2001 at 08:00:23AM -0700, Andy wrote:
> 
> When using the "log from hostname" function in syslog.conf like:-
> 
> #+hostname
> local7.=info  <tab> <tab>  /var/log/file
> #+*
> 
> fails to log to file correctly, in fact, logs nothing.
> 
Works nice here without your patch, modulo the bug David mentioned.
If the sending host is within the same domain, using FQDN for the
sending host does not work, unless the following patch is applied.

I must admit I don't like how the ``#+'' is currently implemented.
For example, it is unclear (and undocumented) from the reading
of the manpage, that the -n flag affects the hostname specification
in syslog.conf(5).


Index: syslogd.c
===================================================================
RCS file: /home/ncvs/src/usr.sbin/syslogd/syslogd.c,v
retrieving revision 1.78
diff -u -p -r1.78 syslogd.c
--- syslogd.c	2001/05/21 22:13:55	1.78
+++ syslogd.c	2001/07/03 08:33:54
@@ -1524,8 +1524,12 @@ cfline(line, f, prog, host)
 	/* save hostname if any */
 	if (host && *host == '*')
 		host = NULL;
-	if (host)
+	if (host) {
+		if ((p = strchr(host, '.')) &&
+		    strcasecmp(p + 1, LocalDomain) == 0)
+			*p = '\0';
 		f->f_host = strdup(host);
+	}
 
 	/* save program name if any */
 	if (prog && *prog == '*')


Cheers,
-- 
Ruslan Ermilov		Oracle Developer/DBA,
ru@sunbay.com		Sunbay Software AG,
ru@FreeBSD.org		FreeBSD committer,
+380.652.512.251	Simferopol, Ukraine

http://www.FreeBSD.org	The Power To Serve
http://www.oracle.com	Enabling The Information Age

> Works nice here without your patch, modulo the bug David mentioned.
> If the sending host is within the same domain, using FQDN for the
> sending host does not work, unless the following patch is applied.

> I must admit I don't like how the ``#+'' is currently implemented.
> For example, it is unclear (and undocumented) from the reading
> of the manpage, that the -n flag affects the hostname specification
> in syslog.conf(5).

Yep - I think this could all be documented better in the man page.
I think the "-n" flag and the host selection stuff were added about
the same time and the man page was never updated to take account
of the interaction.

The patch you included would be a reasonable fix, but I think it
may be better to keep the full host name within syslogd until the
log message is written to the log. This is because syslogd trys to
reconstruct the full host name in atleasy one place after it has
truncated it, which seems silly.

	David.

Comment 7 Dag-Erling Smørgrav 2001-09-13 06:11:14 UTC

State Changed
From-To: analyzed->feedback

Fixed in -CURRENT (unless there are other bugs which were masked by this one) 

Comment 8 Dag-Erling Smørgrav 2001-09-13 06:11:14 UTC

Responsible Changed
From-To: dwmalone->des

I'll handle the shakedown and MFC.

Comment 9 Dag-Erling Smørgrav 2001-11-27 13:50:20 UTC

State Changed
From-To: feedback->closed

This was actually MFCed siz weeks ago, but I forgot to close the PR.