Hi Gerrit,
On Tue, 6 Jul 2004, Gerrit Voss wrote:
> > I'm unfortunately not a network expert, but to me it looks like the
> > clusterServer cannot bind to the multicast-port for some reason. Might be
> > the too new'n'shiny 2.6.x kernel? Or is there a newly introduced bug in
> > OpenSG? :) I use the dailybuild from 30.06.2004.
>
> hmm, nothing changed since I was running a 2+1 cluster (only the servers
> used FC 2, either 32 and 64bit) so I do not think it should be a FC 2
> problem. Which kernel are you using, the one that came with FC 2 or a
> newer one ?. Are you using the normal testClusterServer or your own
> version ??
I use my self-compiled 2.6.6 as mentioned in the original mail. :)
The testClusterServer is pretty much the default, except my own
MultiDisplayOffsetWindow class is compiled in.
But I got a little further: somehow I realised that I need to setup a
default route from the slave machines to the pwmaster (running the
client), so my routing table on the slave side looks simething like:
[EMAIL PROTECTED] root]# /sbin/route
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
192.168.1.0 * 255.255.255.0 U 0 0 0 eth0
169.254.0.0 * 255.255.0.0 U 0 0 0 eth0
127.0.0.0 * 255.0.0.0 U 0 0 0 lo
default pwmaster 0.0.0.0 UG 0 0 0 eth0
When I do an "lsof|grep TCP" on the slave I got the info that it actually
listens on a high port:
[EMAIL PROTECTED] root]# lsof |grep TCP
[...]
testClust 5935 edhellon 11u IPv4 13821 TCP pw4:32778
(LISTEN)
When I do a traceroute on my client I got the following:
[...]
connect(13, {sa_family=AF_INET, sin_port=htons(32779),
sin_addr=inet_addr("192.168.1.4")}, 16) = -1 ECONNREFUSED (Connection
refused)
socket(PF_UNIX, SOCK_STREAM, 0) = 14
connect(14, {sa_family=AF_UNIX, path="/var/run/nscd/socket"}, 110) = -1
ENOENT (No such file or directory)
close(14)
[...]
What the hell is this "/var/run/nscd/socket" file? :)
The client displays the following lines with OSG_LOG_LEVEL set to debug:
[...]
INFO: Connect to right2
INFO: send request to:224.245.211.234
INFO: send request to:255.255.255.255
INFO: Found at address pw4:32779
... and then just seems to hang. Possibly an endless loop because of the
refused connection.
OK, one gotcha found: I copied my testClusterServer.cpp from a machine
that has an older OpenSG version installed, and forgot to add
'ChangeList::setReadWriteDefault();'. Now I've added this line,
recompiled, copied the new binary to all the server-machines, and it's
still the same... I've also added this line to my Client.
I'm going mad...
Akos
-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
digital self defense, top technical experts, no vendor pitches,
unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
