NetworkManager-resolv.conf -

[Ed Greshko]

On 11/18/2012 10:35 PM, Timothy Murphy wrote:
> Ed Greshko wrote:
>
>>>> My server is an HP MicroServer running CentOS-6.3 .
>>> My laptop is a ThinkPad T61p running Fedora-17/KDE .
>>> I have shown above the entry I have in /etc/dhcp/dhcpd.conf on the server
>  
>> What is your AP?
> My AP is a Linksys WRT54GL attached to the server;
> I have another router/modem linking the server to the internet.
>
> Nb My problem only arises when I am on the borderline of the AP -
> "iwlist scan" can see the AP, but before the connection was lost
> iperf showed throughput was down to less than 5% of the usual rate.
> I looked at the NM reports in /var/log/messages ,
> and NM seemed to go through a bewildering sequence of states.
>
> It's not important, as it rarely happens and I know what to do;
> but in my view it shows that NM does not "just work" as is often claimed -
> there remain fairly serious problems, and I think it is important
> to leave the network service as a fall-back when NM fails.
>  

I see.  So, you've disabled the DHCP server on the AP and allow the broadcast through to your server to service DHCP requests.  I've not done that before....but it does sound like something interesting to play with as time permits.   I suppose I'd not be so sure it was an issue with NM in this case.  I'd probably do some packet sniffing to make sure the request/response was being processed at the time of failure.  Probably hard to do if it isn't easy to replicate.

-- 
Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning. -- Rick Cook, The Wizardry Compiled