TNS-12518: TNS:listener could not hand
off client connection
TNS-12547: TNS:lost contact
TNS-12560: TNS:protocol adapter error
TNS-00517: Lost contact
IBM/AIX RISC System/6000 Error: 32:
Broken pipe
This
example of ORA-12518 happens on and off.
The reason
ORA-12518 is being throw may be because of
DEDICATED
connections because Oracle 10g has a value
of PROCESSES is defaulted at 150 which can
be lower than necessary in a production
system. Also, in pre-9i MTS, ORA-12518 may
be thrown in SHARED SERVER
because the dispatcher may have reached the
maximum connection value, so it denies all
other.
There are two solutions for ORA-12518
depending on which symptom you may be
experiencing. For the DEDICATED
occurrence of ORA-12518, you would need to
try increasing the PROCESSES
parameter so that it can handle the needed
number of processes. You can ensure
that you have the needed value by monitoring
the listener log for ORA-12518. Also,
note that because the PROCESSES
parameter is static, the database will need
to be bounced.
If you are
experiencing ORA-12518 because of a shared
server issue, you first would need to use
the command below to shutdown the
dispatcher:
SQL>
alter system shutdown immediate 'D001';
Then, add
on new dispatchers:
SQL>
alter system set DISPATCHERS =
'(protocol=tcp)(dispatchers=3)';
Remember
that MOSC offers these syntaxes as
examples, but resolving ORA-12518 requires
you to evaluate the syntax depending on your
dispatcher value in the Spfile.ora or
init.ora files. As you increase
DISPATCERS to resolve ORA-12518 you
should also keep an eye on the shared server
ratio.
Question:
With Oracle
10g using a Windows platform, how do I
resolve ORA-12518?
Answer:
ORA-12518 is
often due to network issues, and in HP-UX,
there are various bugs causing ORA-12518.
In reference
to your platform, there are a few
resolutions that may be available for
ORA-12518:
- Try
looking in the EventLog to check out
network connections and associated
errors, as well as the alert log in the
database. ORA-12518 can also be
thrown along with a database crash.
- Look
in the windows service for your Oracle
service status
-
start--->run--->services.msc
- The
server may be in need of a swap memory
to disk as it runs out of memory space.
Sometimes, Oracle processes can take up
alot of memory. If you think this
is causing ORA-12518 in your case, you
can try this workaround:
-
set parameter to
DIRECT_HANDOFF_TTC_LISTENER=OFF
in listener.ora
-
restart listener
-
You may also want to consider making
he large_pool_size larger.
