Unable to find known Solaris bug (1084772)?

Question

Hi all,
I'm currently facing an issue with my connectivity between SunOS 5.8 with Oracle.
However I've stumbled upon sites (referenced below) that quoted this "known" Solaris Bug which ties in strongly with the problem I am having.
<Solaris only> Problem Description ------------------- You have a JDBC program connecting to the database. This program may run standalone or on a webserver. This program opens many simultaneous connections to the database. The following error occurs on or near the 257th connection: java.sql.SQLException: ORA-12154: TNS:could not resolve service name All of the following symptoms apply: - Solaris OS - Oracle JDBC/OCI driver - ORA-12154 Error - Failure occurs on 257th connection between JDBC client and Oracle server. - Same program works on a non-Solaris OS, exceeding the 257th connection. Solution Description -------------------- This is a known Solaris bug (1084772). Contact Sun for further information and/or possible patch. As a workaround, you can hardcode the full tns description in the JDBC/OCI connection string. This avoids the need to read the tnsnames.ora file. This will not eliminate the problem, but should free up more file descriptors for your use, thus deferring failure until later. jdbc:oracle:oci8:@(description=(address=(host=pepper)(protocol=tcp)(port=1521))(connect_data=(sid=ORCL))) Explanation ----------- The OS file descriptor limit is being reached. Thus, the tnsnames.ora file cannot be read to resolve the tns alias suppled to the connection string. References ---------- http://sunsolve.sun.com bug (1084772) Summary: Although the limit for file descriptors is 1024, you  cannot open more than 256 fd's using stdio.h. (This is not an Oracle bug.) note:94091.1 Example: Identifying Connection String Problems in JDBC bug:777647 When Opening > 253 File Descriptors and Then Connection bug:2060739 ORA-12154 After Creating 220+ Connections Using Thick (OCI8) JDBC Driver </solaris only>
However when I tried to search for this bug on sunsolve I am not able to find this at all.
Could someone please enlighten me? Is this a real bug or is it some hokey poky?
Thanks
Ref:
http://www.orafaq.com/maillist/oracle-l/2003/12/16/1329.htm
http://www.experts-exchange.com/Database/Oracle/Q_21877697.html
Edited by: uGeneo on Apr 9, 2008 8:12 PM
Added ref.

Answer 1

"The following error occurs on or near the 257th connection:"
There may actually be a bug in the vendor software, but I doubt your application should ever have a need for more than a handful of connections much less 257 connections if you are using connection pooling. Therefore I suggest you look at your application and make sure its using connection pooling and that you are
getting a connection, using it, and closing it all within a try/catch/finally block rather than worry about the vendor bug. Here is a rough outline of
using a connection (all within the same function you your sure the finally block is called, where the conneciton is obtained ,used, and close as quickly as possible):
Connection conn=null;
try{
conn= //get connection
}catch( Exception e){
}finally{
//close connection.
Lets suppose you have 100 people using your application and the above code is called if they hit an 'update' button. The chances of all 100 of them hitting
the update button within the same 10 milliseconds or so for the above code to complete is very very small.
Chances are most of them are spending 99.9993% of the time looking at the web page than clicking the update button. My off the wall guess is that only
2 or 3 people out of 100 will ever access the code at the same 10 milliseconds. Therefore your connection pool only needs 3 connections on stand by.