On 10/29/2009 11:43 AM, Michiel van Es wrote:
Is it possible to reconfigure the jabber service and start clean?
The SYN_RECV packages are not ok..
The strange thing is that I can not even make a normal jabber connection
from a client (say pidgin) so I suspect my jabber settings are garbled..
Is it possible if I post some of my configuration here so people can
look at it?
I don't think it's a jabberd issue ... I feel like this is a general
networking issue. I don't think the application would even see a socket
in that SYN_RECV state. Here's a test:
$ strace nc -l 12345
# tcpdump -vnni lo port 12345 &
# nmap -v -sS -p 12345 localhost
nc doesn't see anything... and the tcpdump shows the kernel sees and
replies to the SYN nmap sends:
16:27:56.880966 IP (tos 0x0, ttl 38, id 64160, offset 0, flags [none],
proto TCP (6), length 44) 127.0.0.1.42004 > 127.0.0.1.12345: S, cksum
0xf26f (correct), 1382315266:1382315266(0) win 3072 <mss 1460>
16:27:56.881000 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto
TCP (6), length 44) 127.0.0.1.12345 > 127.0.0.1.42004: S, cksum 0x0791
(correct), 1846857288:1846857288(0) ack 1382315267 win 32792 <mss 16396>
(then nmap kills the fake connection with a RST...)
So I feel like you should examine the network setup...
Another question:
Is it possbile to set the rhnsd on 5 minutes instead of 4 hours?
In case of a failing jabber server, the clients will pull the requests
every 5 minutes?
Can this do any harm?
That should work - we do something like that here, although we set it to
30 minutes (I think). Edit /etc/sysconfig/rhn/rhnsd - the INTERVAL is
in minutes. Also note that rhnsd varies the time a little every 12
intervals to prevent massive concurrent checkins.
Hope to help,
Josh
_______________________________________________
Spacewalk-list mailing list
Spacewalk-list@redhat.com
https://www.redhat.com/mailman/listinfo/spacewalk-list
