Re: Weird issue with a PACKAGE in 11g

  • From: Lothar Flatz <l.flatz@xxxxxxxxxx>
  • To: oracle-l@xxxxxxxxxxxxx
  • Date: Wed, 26 Aug 2015 00:00:03 +0200

Hi,

error message does it say that the package is invalid. It say the package does not exist. That is something else.
Is it possible that the pacakge runs with invoker rights?

Regards

Lothar
On 25.08.2015 23:26, Marcos Colmenares H. wrote:

Hi Raza,

Thats my main issue actually, it never shows up as invalid... it keeps showing up as valid. In the production enviroment we went as far as doing a mass recompile of all objects, but it happened again.

Sounds interesting as a last resort, recompiling all objects every time they are accesed here would be caos.

Best Regards,

Marcos Colmenares H

--

2015-08-25 16:53 GMT-04:30 Raza Siddiqui <raza.siddiqui@xxxxxxxxxx <mailto:raza.siddiqui@xxxxxxxxxx>>:

Marcos

Look at any dependent objects being dropped or altered...that
event will cause a "stored procedure" (package spec, body,
function or procedure) to become invalid.

If memory serves, there's a parameter that can recompile all
DB-stored pl/SQL code to be recompiled every time it's invoked....

Corrections / additions welcome.

Raza


On Aug 25, 2015, at 14:12, "Marcos Colmenares H."
<mcolmenares@xxxxxxxxxxxxxxxxxxxxxx
<mailto:mcolmenares@xxxxxxxxxxxxxxxxxxxxxx>> wrote:

Good Day,

We've been experiencing waht seems to be a bug with a certain
package, but because of the way it presents itself im not sure if
its really a sole occurence.

We have a package named A with an internal procedure called D
(a.d). It works as intended and when we query invalid or not
compiled objects everything seems OK.

Then one of the logs shows us an error message like so

*ORA-06508: PL/SQL: could not find program unit being called**a.d*



when we re-check for invalid objects (SELECT * FROM all_objects
WHERE owner='xxxxx' AND status='INVALID';), everything seems fine
and i see no indications of any error whatsoever, but around a
month ago trying to fix the issue when it first appeared i
decided to re-compile the package just to try if that fixed it,
and it did. The re-compilation works without a hitch when the
error occurs but its a production enviroment and now its
happening on the testing enviroments as well without any
indication as to why and to be honest im runing out of ideas.


Any of you had a similar situation happen to you?



Best Regards,

Marcos Colmenares H

--




--




Other related posts: