The monitoring machine is slow, the webserver is faster (by a factor of
20+, not even counting that the webserver is SMP). So the connect can
actually succeed at once, even when non-blocking.
Cures this type of false alarm:
Detailed text (if any) follows:
-------------------------------
Detail for id:8 -> www.eksempel.no(0)
+ resolving www.eksempel.no succeeded
+ pack_sockaddr_in command succeeded
+ socket command succeeded
+ fcntl command succeeded
- first connect failed : [1] Operation now in progress
Index: phttp.monitor
===================================================================
RCS file: /home/erik/.cvs/mon/mon.d/phttp.monitor,v
retrieving revision 1.10
retrieving revision 1.12
diff -u -r1.10 -r1.12
--- phttp.monitor 2003/10/30 10:43:41 1.10
+++ phttp.monitor 2003/12/02 10:56:33 1.12
@@ -520,7 +520,7 @@
$client{$id}{"begin"} = gettimeofday;
$command = connect($client{$id}{"fhandle"}, $client{$id}{"hispaddr"});
$debugConnection and print "connect : $!\n";
- if ((not defined($command)) and ($! == EINPROGRESS())){
+ if ($command or ((not $command) and ($! == EINPROGRESS()))){
# Good in non blocking context
$client{$id}{"success"} .= "+ first connect succeeded\n";
$onrace{$id}++;
--
Erik I. Bolsų, Triangel Software AS | Skybert AS
Tlf: 712 41 694 Mobil: 915 79 512
_______________________________________________
mon mailing list
[EMAIL PROTECTED]
http://linux.kernel.org/mailman/listinfo/mon
