Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations TouchToneTommy on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

TCp packet Lost

Status
Not open for further replies.

Gloups

IS-IT--Management
Sep 16, 2003
394
FR
Hi all, i have an interesting and strange TCP problem:


- Two aix boxes on different subnets
- connections and all commands passed by xterm are ok
- connections and all commands passed by ftp are ok

But the same command passed by telnet causes tcp packet lost


Example:

- all commands sending back lot of display like ls -alrt in /tmp are freezing and the same command in ftp session is ok.

- a simple ls in / with only some files to display is ok


Using TCPDUMP on the two systems here commes what i can see

Telnet from machine A to machine B
Command ls -alrt /tmp send by machine A
Command recieved by machine B
Data packet send from machine B to machine A
Packet never arrived to machine A
Machine B still try to retransmit few times

Timeout

Machine B send Reset packet to machine A
Reset packet arrived on machine A


It seems to be ok for small sized packet only

Same tcpip config same mtu
mtu adapted for intermediate router and ok
no specific route only default gateway

The same commands send to another system on the same switch and router as machine B are ok.

Only machine B act this way !!!!


Any idea ??
 
Did you check error log on machines involved? I'm thinking low MBUFs...


HTH,

p5wizard
 
Nothing looking like that in errpt.


I have this problem on all my client PCs trying to connect to machine A
 
Are you NIC's all hard set (1000 Full Duplex) Looks like a speed mismatch to me. Also check the switch.

Mike

"Whenever I dwell for any length of time on my own shortcomings, they gradually begin to seem mild, harmless, rather engaging little things, not at all like the staring defects in other people's characters."
 
i already looked at those type of problems but all NICs are set to 100Mb full duplex and autoneg OFF
 
hi,
try to do a "stress ping" from both machine to
their default gateway (so you don't change subnet and
restrict the test).

"stress ping" I mean

ping -s 100 192.168.1.254
ping -s 1000 192.168.1.254
ping -s 10000 192.168.1.254
ping -s 56000 192.168.1.254

(when you send a stright ping, you send packets 56 byte
long: see output of ping; if you increase the packet size,
you stress the card/cable/switch and data is sent fast;
if your channel is defective, the high frequency by which
you send data, evidentiates the malfunctioning).

If you see one of the 2 boxes that at a certain limits
begins to late the answer, check cable.
Try also to force (if it is still possible today) 10mbs:
if the error disappear, change cables or other and reset
to initial speed.

Try also to ping the machines from your PC(ping -l 100 ..), holding in mind the different subnets between your PC and one or both of the riscs.

ciao
vittorio
 
Thakn's victorv i tried your different ping all are ok exept 56000 bytes size wich doesn't reply.

The thing i found strange is that problem only appear with telnet connections.

all others types of connection are ok not depending of the packet size.

ex: ssh xterm ftp
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top