ORA:12570 and TCP

[Roger Lavender]

Yes, I know - this is not an ORACLE Forum, however, TCP is centric to the HPUX OS so I'm exploring all opportunities.

I have clients attempting to connect to an ORACLE9i database (which is running ok and has a listener that is running ok) on a K570, running HPUX 11.0 32b, and receiving error ORA:12570.

During investigation we ran a client trace and found that data was being transported successfully.

The trace provided the following:

................

nttcni: entry
nttcni: trying to connect to socket 168.
nttcni: exit
nttcon: NT layer TCP/IP connection has been established.
nttcon: set TCP_NODELAY on 168
nttcon: exit
nsopen: transport is open

.......

nscon: recving a packet
nsprecv: entry
nsbal: entry
nsbgetfl: entry
nsbgetfl: normal exit
nsmal: entry
nsmal: 44 bytes at 0x10ad060
nsmal: normal exit
nsbal: normal exit
nsprecv: reading from transport...
nttrd: entry
ntt2err: entry
ntt2err: soc 168 error - operation=5, ntresnt[0]=530, ntresnt[1]=53, ntresnt[2]=0
ntt2err: exit
nttrd: exit
nsprecv: transport read error
nsprecv: error exit

Question: Is there some TCP setting that has been dorked?

There have been no changes to any TCP setting on this server in 127 days.

Thanks in advance.

[James A. Donovan]

Are you seeing ORA-3113 as well?

According to Metalink note #1017442.102 your are experiencing network timeouts due to a very busy network.

http://metalink.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=1017442.102

If you see something very close to the following a SQL*Net/Net8 trace file, 9 times out of ten the network has timed-out. The ntt2err SOC error is the real error - all errors appearing after this are merely symptoms or misleading errors.

ntt2err: soc xxxxx error - operation=5, ntresnt[0]=517, ntresnt[1]=54, ntresnt[2]=0

Close on the heals of ntt2err, you may expect to see something like the following. These errors are again misleading or a catch-22. They tell us that the network or server went down when it only timed-out.

- osnqrc: wanted 1 got 0, type 0 -
- osnqper: error from osnqrc -- osnqper: nr err code: 0 -- osnqper: ns main err code: 12547 -
- osnqper: ns (2) err code: 12560 -
- osnqper: nt main err code: 517 -- osnqper: nt (2) err code: 54 -
- osnqper: nt OS err code: 0 osnqer: entry osnqer: incoming err = 12151

A little later on you may see the following: osnqer: returning err = 3113

The dead ringer for the busy network is: ntt2err: soc xxxx - operation=5, ntresnt[0]=517, ntresnt[1]=54, ntresnt[2]=0

This is very consistent.

[Leon Allen]

Is everything configured to use the same port no.? (1521 usually). Anything funny in the /etc/service file?

[Chris Prescott]

You may want to check the size of your $ORACLE_HOME/network/log/listener.log file to see that it's not overly large. I've had some connection issues in the past when this file gets too large.