DHCP client takes too much time to get IP address

[Alessandro FAGLIA]

Hi list.

I have ISC DHCP 3.0.4-13 running on a Debian Etch box. Our LAN is made 
mainly by Windows 2k and XP boxes, which get their addresses from the 
aforementioned box.

Everything was working with no problem, but after we substituted our 
switches with Netgear models (1xGSM 7248 EU at the center of the star 
and 1xGS 748 TEU and 3xGS 724 TEU at the edges, all connected by 1000-SX 
fiber link) some problems started to arise.

More in detail, some clients take much more time to get their address 
from the DHCP box (some minutes), while in the past the process was very 
fast, as expected.
In the syslog I see following messages:
> Oct 11 06:42:08 cnsrv010 named[17966]: client 127.0.0.1#32900: 
> updating zone 'farmol.it/IN': update unsuccessful: CNWKSXP047
> .farmol.it: 'name not in use' prerequisite not satisfied (YXDOMAIN)
> Oct 11 06:42:08 cnsrv010 named[17966]: client 127.0.0.1#32900: 
> updating zone 'farmol.it/IN': deleting rrset at 'CNWKSXP047.f
> armol.it' A
> Oct 11 06:42:08 cnsrv010 named[17966]: client 127.0.0.1#32900: 
> updating zone 'farmol.it/IN': adding an RR at 'CNWKSXP047.far
> mol.it' A
> Oct 11 06:42:08 cnsrv010 dhcpd: Added new forward map from 
> CNWKSXP047.farmol.it to 192.168.1.67
> Oct 11 06:42:08 cnsrv010 named[17966]: client 127.0.0.1#32900: 
> updating zone '168.192.in-addr.arpa/IN': deleting rrset at '6
> 7.1.168.192.in-addr.arpa' PTR
> Oct 11 06:42:08 cnsrv010 named[17966]: client 127.0.0.1#32900: 
> updating zone '168.192.in-addr.arpa/IN': adding an RR at '67.
> 1.168.192.in-addr.arpa' PTR
> Oct 11 06:42:08 cnsrv010 dhcpd: added reverse map from 
> 67.1.168.192.in-addr.arpa. to CNWKSXP047.farmol.it
> Oct 11 06:42:08 cnsrv010 dhcpd: DHCPREQUEST for 192.168.1.67 from 
> 00:16:76:28:ec:34 via bond0
> Oct 11 06:42:08 cnsrv010 dhcpd: DHCPACK on 192.168.1.67 to 
> 00:16:76:28:ec:34 via bond0
> Oct 11 06:42:09 cnsrv010 dhcpd: DHCPDISCOVER from 00:16:76:28:ec:34 
> via bond0
> Oct 11 06:42:09 cnsrv010 dhcpd: DHCPOFFER on 192.168.1.67 to 
> 00:16:76:28:ec:34 via bond0
> Oct 11 06:42:09 cnsrv010 dhcpd: DHCPDISCOVER from 00:16:76:28:ec:34 
> via bond1: network 192.168.2/24: no free leases
> Oct 11 06:42:09 cnsrv010 dhcpd: DHCPREQUEST for 192.168.1.67 
> (192.168.1.1) from 00:16:76:28:ec:34 via bond1: wrong network.
> Oct 11 06:42:09 cnsrv010 dhcpd: DHCPNAK on 192.168.1.67 to 
> 00:16:76:28:ec:34 via bond1

which repeat every second for some minutes before they stop and the 
client gets its address.

The only configuration step I made on new switches was disabling 
internal DHCP servers.

Anybody experienced such problems? Any hint on where to put my hands?


Thanks and regards.
--
Alessandro