Connections freezing.

[Alan Casey]

We are running a client server application, with Oracle on hp-ux 11.0
We have had intermittant problems connecting from some of the clients. Their connections are freezing and we cant seem to find out why.

I want to rule out a problem with the hpux network card, and anything on the hpux server.
Lanadmin shows:
Can anyone see any clues here?

PPA Number = 0
Description = lan0 Hewlett-Packard 10/100 TX Half-Duplex TT
= 1500
Type (value) = ethernet-csmacd(6)
MTU Size = 1500
Speed = 100000000
Station Address = 0x306e1c61d9
Administration Status (value) = up(1)
Operation Status (value) = up(1)
Last Change = 753
Inbound Octets = 340762801
Inbound Unicast Packets = 643606
Inbound Non-Unicast Packets = 2109777
Inbound Discards = 23
Inbound Errors = 0
Inbound Unknown Protocols = 1429457
Outbound Octets = 72772
Outbound Unicast Packets = 1582
Outbound Non-Unicast Packets = 1
Outbound Discards = 0
Outbound Errors = 0
Outbound Queue Length = 0
Specific = 655367

Press to continue


Ethernet-like Statistics Group

Index = 1
Alignment Errors = 0
FCS Errors = 0
Single Collision Frames = 2
Multiple Collision Frames = 0
Deferred Transmissions = 0
Late Collisions = 0
Excessive Collisions = 0
Internal MAC Transmit Errors = 0
Carrier Sense Errors = 0
Frames Too Long = 0
Internal MAC Receive Errors = 0

[G. Vrijhoeven]

Hi,

Did you check the subnet masks on all components?

Gideon

[T G Manikandan]

One thing to notice above is that the machine is running at Half Duplex and 100Mbps.

make sure that the other side is not in the Full Duplex mode.

Make sure that the speed & duplex match at both ends!

[Mohanasundaram_1]

Hi Alan,

As indicated already, check
1) Duplex settings.
2) Speed settings.
3) cable and ports.

if your problem remains then check the possibility of duplicate IP address in your network.

You can format the /var/adm/nettl.LOG* files through netfmt and check it to find out on duplicate IP.

CHeers,
Mohan.

[rick jones · Accepted solution]

While the link shows half-duplex, the statistics show no _late_ collisions, so the liklihood of there being a duplex mismatch at the HP-UX end is remote. All things considered, those statistics look reasonable.

If the hangs are for specific clients, and you think it is something other than those clients trying to do exotic queries of the database, I would try looking closer to the clients, networking-wise at least.

Might look at netstat -p tcp over an interval:

netstat -p tcp > before; sleep ; netstat -p tcp > after
beforeafter before after > delta

and see what the retransmission rate (overall) might be like. beforeafter comes from ftp://ftp.cup.hp.com/dist/networking/tools/