When I copy/paste an existing package, it inherits all the IDs from the
exisiting package. I generated a new ID on package level for the copied
package (but other level objects still share the same ID ), then I made some
modification on the new package, and let two packages run in parallel in a
master package, which caused tasks cross-run each other. Is this because
those tasks share the same IDs? If so, what' s the easiest way to make the
pasted package have different IDs on all objects from the existing one?

Thanks,

Hello yongli,

Are you saying that after having changed the PackageID for one of the packages
you had issues? What were those issues exactly? How did they cross over?




Allan Mitchell
www.SQLDTS.com
www.SQLIS.com
www.Konesans.com

Hi Allan,

I did change the package ID. The reason I copied/pasted an existing package
to create a new one is because they have the same steps except for details.

I saw this oddity in the debug mode. Although I changed the package IDs,
but the other objects still share the same ID between two packages (I am
wondering if this is the root cause). Both Packages run some Data Flow Tasks
and SQL validations.
The first one did 140 rows, and the second one did 7 rows. This is what I
observed:
when I put both packages in a master package and let them run in parallel
(later I found putting a dependency between them still cause the problem),
The 7 rows package finish successful in Green in Master package while the
other is still running some DFT task, I can see the correponsding task in
yellow in the successful package, kind of odd?
At end the other package failed in red in Master Package, the corresponding
tasks in the successful package changed color to red too even it is green in
the master package, also the row count changed from 7 to 140 something.
It seems to me this is some synchroniztion problem in display in the debug
mode.
(maybe due to those tasks with same IDs).

Although I checked the results to be fine, but this will affect my overal
ETL flow in the master package. This is why I asked if we need to change all
the object IDs in this case?

Thanks.


"Allan Mitchell" wrote:

Hello yongli,


I do not think that this will be an issue. new Package IDs should be sufficient.


Allan Mitchell
www.SQLDTS.com
www.SQLIS.com
www.Konesans.com

One more thing to add in my case. I have a variable defined in both package,
when runing separately, it is fine. When running in paralle, one of the
later finished package failed due to follwoing reason:

Error: The script threw an exception: The element cannot be found in a
collection. This error happens when you try to retrieve an element from a
collection on a container during execution of the package and the element is
not there.

"yongli" wrote:

Hello yongli,

The error suggests that you are trying to ref an object that the task knows
nothing about. Is it a script task that throws the error? How do you let
the task know about the variable? What is the code you use?


Allan Mitchell
www.SQLDTS.com
www.SQLIS.com
www.Konesans.com

I used script task component and specified the variable in package level and
ReadOnlyVariables property in the script task Editor. Since I copied and
pasted the packages so the script task ID remained same in both packages, as
I said I could run each package individaully without problem, but only when I
ran them in parallel, it got the error. This behavior is consistent with the
colorring changing problem described earlier when I ran the pacakges in
parallel, that is SSIS tends to display the object with same ID in same
color, this might be the pitfall in copying/pasting a SSIS package.




"Allan Mitchell" wrote: