NM problems on fedora core 3

[Ivan Blagdan]

Greetings to the list!
First of all i want to say that NM is really a great piece of software, 
that IMHO, does the right thing in getting linux closer to the average 
user. I searched for a decent way of managing connections to the ever 
growing number of APs that i might user with my laptop, and of all the 
numerous apps Linux has to offer, NM looks to be the right way for me, a 
proper example how things should be designed to work on a OS that aims 
to be userfriendly.

Having said that, having it work in most situations would be a welcomed 
thing :) Here is my problem:

Currently i use neat to connect to my home AP trough a configured eth1 
interface, using wep... no problems connecting.
Using NetworkManager to get a connection doesn't work, connection dosn't 
go past the 'connectiong to wireless network' part. Having looked trough 
logs it looks like NM is not getting a DHCP response. Ill put up some 
log entries below.

Ill just say thanks and keep up the good work now, in case you dont make 
it to the end of the message :)

Im running a 2.6.10-1.770_FC3 kernel on Fedora core 3, humming away ever 
so happily on a older Gericom Overdose 3 (sis630 chipset) laptop. I 
managed to get my WLAN adapter running, Gigaset USBadapter 54, using 
ndiswrapper 1.1-1. The AP is Gigaset SE515dsl, a dsl modem/router/ap, 
just like the majority of APs i want to connect to. The AP is configured 
to allow/block access based on client MAC adress. Im mentioning the AP 
explicitly because once i got an connection with NM to a network of the 
WLAN service provider that was around at the time, i could see the 
signal was weak so i didnt poke around much, altough i could ping their 
hosts and access them trough web.

I installed network manager trough yum, here are the package versions:
NetworkManager-0.3.4-1.1.0.fc3.i386.rpm
NetworkManager-gnome-0.3.4-1.1.0.fc3.i386.rpm
Here are some log entries from the time NM starts up (using 
/sbin/service NetworkManager start as root and starting NetworkManager 
as a regular user) untill it tries to connect to the AP again.
Mar 29 18:38:51 progeny NetworkManager: starting...
Mar 29 18:38:51 progeny NetworkManager: NetworkManager startup succeeded
Mar 29 18:38:53 progeny NetworkManager: eth0: Driver support level for 
'sis900' is fully-supported
Mar 29 18:38:53 progeny NetworkManager: nm_device_new(): waiting for 
device's worker thread to start.
Mar 29 18:38:54 progeny NetworkManager: nm_device_new(): device's worker 
thread started, continuing.
Mar 29 18:38:54 progeny NetworkManager: Now managing wired device 'eth0'.
Mar 29 18:38:55 progeny NetworkManager: eth1: Driver support level for 
'ndiswrapper' is fully-supported
Mar 29 18:38:58 progeny NetworkManager: nm_device_new(): waiting for 
device's worker thread to start.
Mar 29 18:38:59 progeny NetworkManager: nm_device_new(): device's worker 
thread started, continuing.
Mar 29 18:38:59 progeny NetworkManager: Now managing wireless device 'eth1'.
Mar 29 18:38:59 progeny nscd: nscd shutdown failed
Mar 29 18:38:59 progeny nscd: nscd startup succeeded
Mar 29 18:38:59 progeny NetworkManager: Clearing nscd hosts cache.
Mar 29 18:38:59 progeny NetworkManager: SWITCH: best device changed
Mar 29 18:39:00 progeny nscd: 15232 Access Vector Cache (AVC) started
Mar 29 18:39:02 progeny NetworkManager: Activation (eth1) started...
Mar 29 18:39:03 progeny NetworkManager: Activation (eth1/wireless): 
waiting for an access point.
Mar 29 18:39:10 progeny su(pam_unix)[10794]: session closed for user root
Mar 29 18:39:17 progeny NetworkManager: Updating allowed wireless 
network lists.
Mar 29 18:39:19 progeny NetworkManager: Activation (eth1/wireless): 
found access point 'done_wlan' to use.
Mar 29 18:39:19 progeny NetworkManager: Activation (eth1/wireless): 
access point 'done_wlan' is encrypted, and a key exists.  No new key needed.
Mar 29 18:39:24 progeny NetworkManager: Activation (eth1/wireless): 
using essid 'done_wlan', with Open System authentication.
Mar 29 18:39:27 progeny nmbd[4055]: [2005/03/29 18:39:27, 0] 
libsmb/nmblib.c:send_udp(790)
Mar 29 18:39:28 progeny nmbd[4055]:   Packet send failed to 
192.168.1.255(137) ERRNO=Network is unreachable
Mar 29 18:39:28 progeny nmbd[4055]: [2005/03/29 18:39:28, 0] 
nmbd/nmbd_packets.c:send_netbios_packet(163)
Mar 29 18:39:28 progeny nmbd[4055]:   send_netbios_packet: send_packet() 
to IP 192.168.1.255 port 137 failed
Mar 29 18:39:28 progeny nmbd[4055]: [2005/03/29 18:39:28, 0] 
nmbd/nmbd_namequery.c:query_name(237)
Mar 29 18:39:28 progeny nmbd[4055]:   query_name: Failed to send packet 
trying to query name PRESUDST1d
Mar 29 18:39:28 progeny nmbd[4055]: [2005/03/29 18:39:28, 0] 
nmbd/nmbd.c:reload_interfaces(265)
Mar 29 18:39:28 progeny nmbd[4055]:   reload_interfaces: No subnets to 
listen to. Shutting down...
Mar 29 18:39:31 progeny NetworkManager: dhcp_interface_init: MAC address 
= 00:30:f1:dd:e0:1c
Mar 29 18:39:31 progeny NetworkManager: Broadcasting DHCP_DISCOVER
Mar 29 18:39:31 progeny 

Re: NM problems on fedora core 3

[Ivan Blagdan]

Ivan Blagdan wrote:
Currently i use neat to connect to my home AP trough a configured eth1 
interface, using wep... no problems connecting.
Using NetworkManager to get a connection doesn't work, connection 
dosn't go past the 'connectiong to wireless network' part. Having 
looked trough logs it looks like NM is not getting a DHCP response. 
Ill put up some log entries below.
I forgot to add one more detail. NM handles the wired connection to the 
same AP/router trough eth0.

Sorry for my english :)
Cheers,
   Ivan
___
NetworkManager-list mailing list
NetworkManager-list@gnome.org
http://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: NM problems on fedora core 3

[Bill Moss]

Your scan shows two AP's, yours with a signal level of -72 dBm and 
another one with signal level of -77 dBm. These are both weak signals 
and I expect if you 'watch -n2 iwlist eth1 scan' you will see these 
levels fluctuate a good deal. Both AP's are set to the same channel 11. 
This will cause interference. I would change the channel of your AP to 1 
or 6 and see what happens. Two AP's cannot operate on the same channel 
in the same location.

--
Bill Moss
Professor, Mathematical Sciences
Clemson University
___
NetworkManager-list mailing list
NetworkManager-list@gnome.org
http://mail.gnome.org/mailman/listinfo/networkmanager-list

Re: NM problems on fedora core 3

[Ivan Blagdan]

Bill Moss wrote:
Your scan shows two AP's, yours with a signal level of -72 dBm and 
another one with signal level of -77 dBm. These are both weak signals 
and I expect if you 'watch -n2 iwlist eth1 scan' you will see these 
levels fluctuate a good deal. Both AP's are set to the same channel 
11. This will cause interference. I would change the channel of your 
AP to 1 or 6 and see what happens. Two AP's cannot operate on the same 
channel in the same location.

Thas all true and fine right now, im not sure my card is reporting 
signal levels right anyway. I reconfigured my AP for channel 6 all the 
same; NM doesn't change its behaviour (no DHCP response). My main 
concern is that NM fails to connect whereas neat succeeds regardless of 
the chan and/or signal level the card reports.

Thanks for the suggestion,
   Ivan
___
NetworkManager-list mailing list
NetworkManager-list@gnome.org
http://mail.gnome.org/mailman/listinfo/networkmanager-list