Hello,
I have some in-house software running under LINUX that does a tcp connect to
a process (port # 32659) running on an HP A900 (RTE-A OS with a NS1000
tcp/ip package addon- yes I know this is a dinosaur but it's not mine). The
connect and communications worked fine on Redhat 4.2 Kernel 2.0.30, but when
I upgraded to Redhat 6.0 and 6.1 the connect no longer works (ping does
work). Netstat show the port is in the SYN_SENT state and it stays that way
until it times out. tcpdump shows the initial connect request going out a
few times followed by an arp request which gets answered and it retries
again (I never see any incoming messages). The LAN analyzer shows the
request going out and a response coming back from the processor (tcpdump
never shows this response).
If I boot the Redhat 4.2 software with the same hardware the connection is
immediate. The data on the LAN analyzer is different from the RH6.1 boot.
The RH4.2 software multicasts out a message before the connect msg goes out,
also the connect response from the processor appears different.
I have tried turning off various things in the kernel (FIREWALL). I also
noticed the /proc/sys/net options but was not sure what to play with there.
The processor is on the same network as I am connected via hub (no routing).
I am using a Pentium III 450 with an onboard EtherExpress 100 lan using the
eepro100 driver.
Unfortunately this old processor will not be replaced for another year and I
am forced to get this working. Any help will be greatly appreciated.
Thanks
Rhonda Moore
Raytheon Co.- Aircraft Integration Systems
903.457.6676
903.457-4413 (fax)
moorera@gvl.esys.com
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.rutgers.edu
This archive was generated by hypermail 2b29 : Tue Feb 29 2000 - 21:00:26 EST