From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.2) Gecko/20031013 Description of problem: The satellite installer contains a procedure, checkDatabaseVersion(), which runs the following PL/SQL code: begin dbms_utility.db_version(:v,:c); end; The bind variables, v and c are declared as 10 char long strings, however this is not long enough as the Database verison on Oracle 9iR2 on Solaris returns as 9.2.0.3.0.0 which is 11 chars. This causes the installer to bail with an error. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.Install RHN satellite connected to an instance of 9iR2 running on a Solaris box. 2. Proceed to db initialisation/schema creation. 3. Actual Results: Immediately after certificate checking, an error window displays: ORA-06502: PL/SQL: numeric or value error: character string buffer too small ORA-06512: at "SYS.DBMS_UTILITY". line 519 ORA-06512: at line 1 {'SQL':'begin dbms_utility.db_version( :v,:c );end;','Params':[(':c',''),(':v','')]}) Expected Results: Installer should have created schema. Additional info: Patch: cd install rm satDB.pyc Apply following patch: --- satDB.py~ Mon Dec 1 15:10:27 2003 +++ satDB.py Mon Dec 1 12:34:14 2003 @@ -91,8 +91,8 @@ def checkDatabaseVersion(): "Current supported Oracle DB versions are 8.1.7 and 9.2.0" - db_ver = ' ' * 10 - db_compat = ' ' * 10 + db_ver = ' ' * 20 + db_compat = ' ' * 20 h = prepare("begin dbms_utility.db_version( :v, :c ); end;"); h.execute(v = db_ver, c = db_compat) version = string.join(string.split(db_ver,'.')[:3],'')
Lovely! Appling patch. Thanks. TEST PLAN: ---------- o hmm.... difficult to test. We'll have to set up an environment that reports this version. o ensure embedded DB and our DB's still report versions correctly.
Need a funky db - over to Misa, who will be back in a month. Will verify it does not break existing code until then.
I don't feel like re-installing a full instance just for this. We can probably hack the query to return a very long string instead of the real versions.
komejo/jamo : please proceed with setting up the environment for testing this bug (i.e. have a satellite installed). Then poke me and I'll slightly change the code to return a longer string than usual.
There doesn't seem to have been any activity on this for 8 months. Is it fixed? Can I tell my client that he can remove the hack I did for him?
Moving to NEEDINFO_ENG - he'll never see it ever otherwise.
NEEDINFO_ENG has been deprecated in favor of NEEDINFO or ASSIGNED. Changing status to ASSIGNED for ENG review.
This has been long fixed. db_ver = ' ' * 50 db_compat = ' ' * 50