[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: DHCP & IP ranges (10.0.2.0/22?)



On Saturday 28 November 2009 14:16 RalfGesellensetter wrote:
> Dear list,
>
> I'd like to ask two (connected) questions:
>
> the current configuration allows over 500 clients (given
> 10.0.2.0/23), however I wonder if it was a big deal to just double
> that range by making it 10.0.2.0/22?
>
> Furthermore, we have a lab that happens to use up twice the amount of
> IP addresse, because (now!):
>
>  * every computer has got _two_ network cards
>  * additional to onboard nic (not supported by earlier kernels) there
>    is a pci nic (3com) in every computer.
>  * students (and probably also teachers) cannot be kept from
>    un/re-plugging cables, so you never know what nic is connected.
>  * each computer, hence, got 2 mac addresses (used exclusively!)
>
> Now, I wonder if it is possible to assign each IP twice in dhcpd.conf
> (once for each network card, because those are used exclusively).
>
> Thank for any clarification.

Making logical and physical networks become as big as described here, 
forces next problems coming up.
Network browsing windows clients not finding their shares are one of 
them. Also the NFS using clients mounting their homes and all the other 
amount of network traffic that will be produced then, will cause a 
single tjener to collapse if not dimensioned properly.
So, if it is really wanted, I think it should be better to split 
services onto specialized machines to not overload the network and the 
tjener.
On the other hand, most of the installations I know, are very small 
ones. And I think, the todays network layout and services structure is 
even overblown for such small schools I know with around 50 clients and 
one server. The people which realy need such amounts of clients should 
better do that by making it by their own or hire professional support 
helping them.

And last but not least, what prevents you from switching the onboard 
nic's off in the BIOS. And furthermore I recommend you not to try to 
solve human problems with technical solutions, because it will not 
work.

Just my two cents.

Jürgen Leibner

-- 
never change a running system
change a never running system
run a never changing system


Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: