problem of packet loss with e100 machine with fedora core 4

[Eric Doutreleau]

Bob Chiodini wrote:

>On Thu, 2005-12-22 at 14:54 +0100, Eric Doutreleau wrote: 
>  
>
>>Bob Chiodini wrote:
>>
>>    
>>
>>>On Thu, 2005-12-22 at 12:11 +0100, Eric Doutreleau wrote:
>>> 
>>>
>>>      
>>>
>>>>Tim wrote:
>>>>
>>>>   
>>>>
>>>>        
>>>>
>>>>>On Wed, 2005-12-21 at 16:11 +0100, Eric Doutreleau wrote:
>>>>>
>>>>>
>>>>>     
>>>>>
>>>>>          
>>>>>
>>>>>>i have a lab with fedora core 4 machine with a network card 
>>>>>>01:08.0 Ethernet controller: Intel Corporation 82562EZ 10/100 Ethernet
>>>>>>Controller (rev 02)
>>>>>>
>>>>>>and from time to time the network seems not working anymore.
>>>>>>
>>>>>>after a few second the network goes up again.
>>>>>>
>>>>>>The result is that i lose about 30% of packet.
>>>>>>
>>>>>>when i boot windows on the same machine i have the same problem.
>>>>>>  
>>>>>>
>>>>>>       
>>>>>>
>>>>>>            
>>>>>>
>>>>>That does seem to suggest a hardware fault (a bad NIC, or the
>>>>>combination of your motherboard and NIC).  Can you try a different one?
>>>>>
>>>>>
>>>>>
>>>>>     
>>>>>
>>>>>          
>>>>>
>>>>well i made a mistake with my sentence
>>>>i have no problem under windows
>>>>and i have 12 same computer and all have the problems with FC4
>>>>
>>>>nevertheless thanks for your answer
>>>>
>>>>   
>>>>
>>>>        
>>>>
>>>Eric,
>>>
>>>What does ethtool eth0 tell you about the interface?  Are you getting
>>>any errors in /var/log/messages concerning eth0?
>>>
>>>This sounds like an autonegotiation (duplex) problem.
>>>
>>>Bob...
>>>
>>> 
>>>
>>>      
>>>
>>here is the result
>>Settings for eth0:
>>        Supported ports: [ TP MII ]
>>        Supported link modes:   10baseT/Half 10baseT/Full
>>                                100baseT/Half 100baseT/Full
>>        Supports auto-negotiation: Yes
>>        Advertised link modes:  10baseT/Half 10baseT/Full
>>                                100baseT/Half 100baseT/Full
>>        Advertised auto-negotiation: Yes
>>        Speed: 100Mb/s
>>        Duplex: Full
>>        Port: MII
>>        PHYAD: 1
>>        Transceiver: internal
>>        Auto-negotiation: on
>>        Supports Wake-on: g
>>        Wake-on: g
>>        Current message level: 0x00000007 (7)
>>        Link detected: yes
>>
>>the only thing i got from messages concerning eth0 is theses messages
>>
>>Dec 22 12:17:04 b02-07 kernel: e100: eth0: e100_probe: addr 0xfc000000, 
>>irq 11, MAC addr 00:0C:76:E5:E8:C8
>>Dec 22 12:17:08 b02-07 kernel: e100: eth0: e100_watchdog: link up, 
>>100Mbps, full-duplex
>>Dec 22 12:17:15 b02-07 ntpd[2459]: Listening on interface eth0, 
>>157.159.15.231#123
>>
>>
>>    
>>
>
>Eric,
>
>Presumably those messages are from a boot up?  Do you get the same ping
>failures pinging something attached to the same switch?
>  
>

Yes i have the same problems between machines from the same switch
i test between two linux machine and one machine under windows and the other
with linux

>What does cat /proc/sys/net/ipv4/tcp_ecn return?  If it's not zero, set
>it to zero echo 0 > /proc/sys/net/ipv4/tcp_ecn.
>  
>

it s zero

>I have no trouble pinging you at 157.159.15.226, 227, 228, 232, and 234.
>Zero packet loss.  The others do not respond at all.
>
>  
>

well as it s an open lab all the machine you told were rebooted under 
windows.

>Do you get the same results using ping instead of fping?
>
>  
>

i just try with ping and i got the following

94 packets transmitted, 58 received, 38% packet loss, time 93222ms
rtt min/avg/max/mdev = 0.247/0.271/0.294/0.017 ms, pipe 2

then same problem
the machine sporadicly stop to answer to the ping :(

thanks for your answer

>Bob...
>
>Bob...
>
>Bob...
>
>
>
>
>  
>