[PLUG] Baffled: CentOS 7 DHCP 10G fast, fixed 1G

Louis Kowolowski louisk at cryptomonkeys.org
Thu Nov 14 15:54:12 UTC 2019


Agreed. If there are ways to inspect the Forwarding Information Base, thats where you will find where next-hop traffic is sent. Linux may call it something else. FIB is a term from the network world. In essence, its a map of MAC addresses and interfaces, with (typically) 1 being the default.


> On Nov 14, 2019, at 8:47 AM, Tomas Kuchta <tomas.kuchta.lists at gmail.com> wrote:
> 
> It seems that you are on the way to discover that the 10gb link is fine,
> but your bits go through the 1gb interface.
> 
> If that is the case, you have some decisions to make about network
> config/routing. You can check how your DHCP client avoids it, perhaps it
> simply let's the 1gb interface unassigned, or it is by order...
> 
> Tomas
> 
> 
> On Thu, Nov 14, 2019, 06:52 Michael Dexter <dexter at pdxlinux.org> wrote:
> 
>> On 11/13/19 1:07 PM, alan at clueserver.org wrote:
>>> What does ethtool tell you?
>> 
>> "DHCP" via network-scripts:
>> 
>>  Settings for p2p1:
>> Supported ports: [ TP ]
>> Supported link modes:   100baseT/Full
>>                         1000baseT/Full
>>                         10000baseT/Full
>> Supported pause frame use: Symmetric
>> Supports auto-negotiation: Yes
>> Advertised link modes:  100baseT/Full
>>                         1000baseT/Full
>>                         10000baseT/Full
>> Advertised pause frame use: Symmetric
>> Advertised auto-negotiation: Yes
>> Speed: 10000Mb/s
>> Duplex: Full
>> Port: Twisted Pair
>> PHYAD: 0
>> Transceiver: external
>> Auto-negotiation: on
>> MDI-X: Unknown
>> Supports Wake-on: d
>> Wake-on: d
>> Current message level: 0x00000007 (7)
>>       drv probe link
>> Link detected: yes
>> 
>> iperf3 sample:
>> 
>> [  4]   0.00-1.00   sec   114 MBytes   955 Mbits/sec    0    375 KBytes
>> 
>> [  4]   1.00-2.00   sec   113 MBytes   946 Mbits/sec    0    375 KBytes
>> 
>> 
>> 
>> $ sudo dhclient p2p1
>> [cecheverri at rbx-1 ~]$ sudo ethtool p2p1
>> Settings for p2p1:
>> Supported ports: [ TP ]
>> Supported link modes:   100baseT/Full
>>                         1000baseT/Full
>>                         10000baseT/Full
>> Supported pause frame use: Symmetric
>> Supports auto-negotiation: Yes
>> Advertised link modes:  100baseT/Full
>>                         1000baseT/Full
>>                         10000baseT/Full
>> Advertised pause frame use: Symmetric
>> Advertised auto-negotiation: Yes
>> Speed: 10000Mb/s
>> Duplex: Full
>> Port: Twisted Pair
>> PHYAD: 0
>> Transceiver: external
>> Auto-negotiation: on
>> MDI-X: Unknown
>> Supports Wake-on: d
>> Wake-on: d
>> Current message level: 0x00000007 (7)
>>       drv probe link
>> Link detected: yes
>> 
>> iperf3 sample:
>> [  4]   0.00-1.00   sec  1.10 GBytes  9.41 Gbits/sec    0   1.23 MBytes
>> 
>> [  4]   1.00-2.00   sec  1.10 GBytes  9.42 Gbits/sec    3   1.14 MBytes
>> 
>> [  4]   2.00-3.00   sec  1.09 GBytes  9.40 Gbits/sec    3   1.14 MBytes
>> 
>> 
>> So, identical ethtool output unless I'm missing something?
>> 
>> Maybe responsible: There are two NICs on the same subnet as not to get
>> locked out. Some OSs tolerate this while some do not. Because "dhclient"
>> addressing works fine, I trust that is okay but the plan is to next try
>> it without the 1GbE interface connected.
>> 
>> Other ideas?
>> 
>> In short, 'dhclient <interface>' gives a solid 10GbE. DHCP via
>> network-scripts drops to 1GbE.
>> 
>> Michael
>> _______________________________________________
>> PLUG mailing list
>> PLUG at pdxlinux.org
>> http://lists.pdxlinux.org/mailman/listinfo/plug
>> 
> _______________________________________________
> PLUG mailing list
> PLUG at pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug

--
Louis Kowolowski                                louisk at cryptomonkeys.org <mailto:louisk at cryptomonkeys.org>
Cryptomonkeys:                                   http://www.cryptomonkeys.com/ <http://www.cryptomonkeys.com/>

Making life more interesting for people since 1977




More information about the PLUG mailing list