Running Oracle 9.2 on a Solaris machine, I get the below error nearly
each day.
Restarting the oracle listener helps, else after a while we are
getting TNS-12518: TNS:listener could not hand off client connection
errors. So new connections are either hard or impossible to get.
This is part of the listening trace.
...
nttgetport: port resolved to 59357
nttbnd2addr: looking up IP addr for host: charxxxxx.xxx
ntus2err: sd=113, op=14, resnt[0]=507, resnt[1]=0, resnt[2]=0
nsbequeath: error reading REDIR/NSE msg
nserror: nsres: id=101, op=72, ns=12560, ns2=0; nt[0]=0, nt[1]=0,
nt[2]=0; ora[0
nstimarmed: no timer allocated
nsclose: closing transport
...
listener.log
29-JUN-2004 11:08:11 *
(CONNECT_DATA=(SID=imx)(CID=(PROGRAM=)(HOST=haxxx)(USER=a
pache))) * (ADDRESS=(PROTOCOL=tcp)(HOST=194.x.x.x)(PORT=44829)) *
establish
* imx * 12518
TNS-12518: TNS:listener could not hand off client connection
TNS-12560: TNS:protocol adapter error
...
The number of registered connections is raising all the time after
some time.
nsevwait: entry
nsevwait: 663 registered connection(s)
nsevwait: 0 pre-posted event(s)
nsevwait: waiting for transport event (1 thru 664)...
nsevwait: 1 newly-posted event(s)
nsevfnt: cxd: 0x1430b10 stage 4: NT events set:
WRITE
nsevfnt: cxd: 0x1430b10 stage 4: NS events set:
OUTGOING CALL COMPLETE
....
Any ideas ?
David.
> ntus2err: sd=113, op=14, resnt[0]=507, resnt[1]=0, resnt[2]=0
> nsbequeath: error reading REDIR/NSE msg
> nserror: nsres: id=101, op=72, ns=12560, ns2=0; nt[0]=0, nt[1]=0,
> nt[2]=0; ora[0
You are having REDIR/NSE messages, which means that you're hitting
a router of some kind. Do normal network connections like telnet,
ssh, scp,ftp, or X11 work normally? Protocol error usually signifies
either library mismatch or a network problem.
--
Well-behaved women seldom make history
I'm not sure whether you are right. The bequeath protocol after all IS
non-routable.
--
Sybrand Bakker, Senior Oracle DBA
>
> I'm not sure whether you are right. The bequeath protocol after all IS
> non-routable.
You are right, it's BEQ protocol. Well then, if netca doesn't do the trick,
there is a problem with software.
in linux, you can set the range of ports to be used in /etc/sysctl
use sysctl -a to view the settings.
sorry, I'm still relatively solaris illiterate, and my Solaris ref
manuals are on the bookshelf at work.
why not try reducing the range of ports to say 1024::49000 and give it
a go again? you may be bumping into NAT, ip_masq or some other pothole
up in the high port range, perhaps a trojan ? (on solaris, yeah,
right)
Pd
in linux, you can set the range of ports to be used in /etc/sysctl