Oracle Listener not responding

From: <waimanfung72_at_gmail.com>
Date: 28 Apr 2005 20:17:52 -0700
Message-ID: <1114744672.854983.190920@g14g2000cwa.googlegroups.com>

Hi,

Thanks in advance.

We have a situation where a machine is streaming large amounts of data via multiple oracle connections to our Oracle DB. Over time, the oracle listener does not respond to any more connection requests. Only after 10 minutes of inactivity, the oracle listener begins to listen to new requests. This scenario is easily reproducible in our environment.

Below is the log for the listener when the oracle listener stops listening to new requests: (Normally the log woulf finish with an "sntpcall: result string is NTP0 xxxx" but it does not have this entry when it happens)

sntpcall: hdl[IR]=19, hdl[IW]=18
nsbeqaddr: doing connect handshake...
nsbequeath: doing connect handshake...
nsbequeath: NSE=0
nsbeqaddr: connect handshake is complete nstimarmed: no timer allocated

nsclose: closing transport
nsclose: global context check-out (from slot 5) complete
tnslsnr: connect string inst->lcxd[0]:

(ADDRESS=(PROTOCOL=tcp)(HOST=sdc-db001)(PORT=1521))

nsopen: opening transport...
nttcnp: getting sockname
nttcnr: waiting to accept a connection.
nttcnr: getting sockname
nttcon: set TCP_NODELAY on 16
nsopen: transport is open

nsnainit: inf->nsinfflg[0]: 0xd inf->nsinfflg[1]: 0xd nsopen: global context check-in (to slot 5) complete nsanswer: deferring connect attempt; at stage 5 nscon: doing connect handshake...
nscon: got NSPTCN packet
nsevdansw: exit
nsbeqaddr: connecting...
sntpcall: detaching from parent with additional fork

We are looking at increasing the oracle queue size (8.1.7) and incresing the allowable TCP connections (tcp_conn_req_max_q) on the OS level (Sun Solaris).

Much appreciated if anyone could provide further assistance.

Thanks.

Regards,
Ray Received on Thu Apr 28 2005 - 22:17:52 CDT