Re: Specifying a JVM for JDBCR4, was Re: a problem accessing a MySQL server

James,

You can easily capture stdout in an interactive job by overriding it prior to starting the JVM. So at the beginning of the job (before the JVM has been started) do this:

OVRPRTF FILE(STDOUT) TOFILE(QSYSPRT) HOLD(*YES)

Then the JVM will write stdout to a spooled file. The spooled file won't end until the job ends, though... but once it does, all the info will be there.

Not sure why using Java 6 would make any difference? Nothing you're doing requires 6.

The error "incorrect database name ''" means that you have not specified a database name. Or, at least, that's what it seems to be saying, right? The database name is ''

MySQL does not have any notion of schema. It just has a database name. DB2 for i is almost a complete opposite in that respect. It uses schema for everything, and there's only one actual database on the whole system, so you almost never select a database. By contrast, MySQL does not have schemas, but can have as many databases as you like, so people tend to use databases similarly to the way we use schemas.

Anyway... Are you selecting a database, either in your connection properties, or via the "use" command? IF not, that would explain the error.

-SK



On 2/20/2013 6:09 PM, James H. H. Lampert wrote:

On 2/20/13 3:10 PM, Scott Klement wrote:

ADDENVVAR ENVVAR(JAVA_HOME)
VALUE('/QOpenSys/QIBM/ProdData/JavaVM/jdk50/32bit')

Thanks.

I was able to successfully tell it to use Java 6. And there was a change
in the behavior.

Unfortunately, not a very useful change: the stuff sent to STDOUT
changed a bit, but it simply sat there for even longer, chewing on
itself, before it finally started emitting multiple dumps. And the
exception reported in the joblog is still:

Java exception "com.mysql.jdbc.exceptions.jdbc4.MySQLSyntaxErrorException:
Incorrect database name ''" when calling method "getPrimaryKeys" with
signature
"(Ljava.lang.String;Ljava.lang.String;Ljava.lang.String;)Ljava.sql.ResultSet
;" in class "java.sql.DatabaseMetaData".

(that's two apostrophes and a double quote, as the "incorrect database
name.)

In an earlier experiment, I was able to determine that changing the case
of the schema made no difference in the error message.

Is there a way, with either the debugger, or some addition to the
service program, or what comes out STDOUT, or the dumps (which don't
actually happen until the JVM crashes trying to free null), that I can
find out what is actually being passed to the Java "getPrimaryKeys" method?

--
JHHL