Issue with PowerEdge 1850 network connectivity

[Matthias Fuchs]

Hi,

did you find a solution?

It seems that we have the same issue with our PE 1850.
We are running Debian Etch with kernel 2.6.15. I recently updated the 
kernel's e1000 driver to 7.0.33 which is the most actual version.

We exchanged networking cabling, moved from a 100 Base-T Switch to 
Gigabit, played around with module parameter an ethtool options...

The PE 1850 is connected to our intranet. I have no problems to connect 
to the PE 1850 from some other Linux boxes. ping, telnet, NFS, TFTP and 
others are fine.

 From several W2K maschines and some other Linux boxes we get ping loss, 
havily delayed telnet session and even telnet sessions being disconnected.

Here is the sniffed output from a W2K client (10.0.18.3) trying to 
telnet to the PE 1850 (10.0.0.190):

       3 0.000321    10.0.18.3             10.0.0.190            TCP 
   1437 > telnet [SYN] Seq=0 Len=0 MSS=1460
       4 0.000705    10.0.0.190            10.0.18.3             TCP 
   telnet > 1437 [SYN, ACK] Seq=0 Ack=1 Win=5840 Len=0 MSS=1460
       5 0.000738    10.0.18.3             10.0.0.190            TCP 
   1437 > telnet [ACK] Seq=1 Ack=1 Win=65535 Len=0
       6 0.007169    Dell_52:a4:6c         Broadcast             ARP 
   Who has 10.0.0.79?  Tell 10.0.0.190
       7 0.007902    10.0.0.190            10.0.18.3             TELNET 
   Telnet Data ...
       8 0.008130    10.0.18.3             10.0.0.190            TELNET 
   Telnet Data ...
       9 0.008620    10.0.0.190            10.0.18.3             TCP 
   telnet > 1437 [ACK] Seq=13 Ack=7 Win=5840 Len=0
      10 0.008634    10.0.0.190            10.0.18.3             TELNET 
   Telnet Data ...
      11 0.008656    10.0.18.3             10.0.0.190            TELNET 
   Telnet Data ...
      12 0.009112    10.0.0.190            10.0.18.3             TELNET 
   Telnet Data ...
      13 0.009133    10.0.18.3             10.0.0.190            TELNET 
   Telnet Data ...
      14 0.016402    Dell_52:a4:6e         Intel_4d:ff:03        ARP 
   10.0.0.190 is at 00:13:72:52:a4:6e
      15 0.049170    10.0.0.190            10.0.18.3             TCP 
   telnet > 1437 [ACK] Seq=28 Ack=25 Win=5840 Len=0
      16 0.049256    10.0.18.3             10.0.0.190            TELNET 
   Telnet Data ...
      17 1.465725    10.0.18.3             10.0.0.190            TELNET 
   [TCP Retransmission] Telnet Data ...
      18 4.470038    10.0.18.3             10.0.0.190            TELNET 
   [TCP Retransmission] Telnet Data ...
      19 10.378557   10.0.18.3             10.0.0.190            TELNET 
   [TCP Retransmission] Telnet Data ...
      20 21.578707   10.0.18.3             10.0.0.190            TCP 
   1437 > telnet [FIN, ACK] Seq=41 Ack=28 Win=65508 Len=0
      21 22.395868   10.0.18.3             10.0.0.190            TELNET 
   [TCP Retransmission] Telnet Data ...

In this case I do not get a prompt of the remote maschine. Please note 
the TCP retransmissions. When running a tcpdump on the other side at the 
same time I can also see these packets on the PE. So they seem to arrive 
on the PE.

Could this be a upper layer issue? I have no idea.
I am also not sure which further information I should report since we 
already change lots of parameters.

Regards
Matthias



JackHammer19 wrote:
> I have a Dell PowerEdge 1850 web server with dual Intel PRO/1000
> network interfaces built into the motherboard.  The operating system is
> CentOS 4.2 (identical to Red Hat Enterprise Linux AS 4.2 - built from
> Red Hat sources).  I am running the latest kernel ( 2.6.9-34.ELsmp).
> The lspci output reports the interfaces as Intel 82541GI/PI. 
> 
> The system has the latest patches applied.