I have followed Jamie's excelent article on logging to a table to follow the execution of a package.
I have a master package; with an OnPostExecute event handler, and a looping sub package that is called from the master. The problem is that ALL steps from the sub package are going through the master's event handler whether the disable event handler is set to true or not. (including any event handler steps in the sub package)
Is this a bug or an anomally for the sub/master relationship
Interestingly enough, when there are Onpostexecute event handlers in both packages the sub steps gets logged twice, once as part of the master package and once as part of the sub.
Paul Pisarek
PS Glad to see Jamie back online...got a bit concerned for him when there were no posts after the London bombings

Disable event handlers not valid for sub packages
fkmfkm
Thanks so much for the sentiment. On a purely personal level I'm happy to be able to say that no-one I know is directly affected by the bombings even though I know alot of people that travel those routes every day. Grateful for small mercies I guess.
As for your SSIS problem...I can replicate what you are seeing. And if you think about it, it makes sense. By setting DisableEventHandlers=TRUE in the child package all you are doing is saying "If an event is raised, don't do anything with it in this package". It is not saying "Don't raise any events". In other words, the events still occur, they just don't get handled (not in the child package anyway).
So its not a bug. Its just something you have to be aware of.
HOWEVER...I was trying a workaround. I figured that attaching an eventhandler that doesnt actually do anything to a task in the child package and then setting System::Propogate=FALSE in that eventhandler would stop the event being handled by the parent package. I was wrong, the event in the child package was still handled by the parent package.
I think this is a bug. I'd be interested to see what Kirk et al have to say about this.
One last thought...how about a property on a package (e.g. PropogateEventsToMaster) that when set to FALSE stopped events getting propogated to parent package(s). Comments
-Jamie
P.S. I have a couple of simple demo packages to...well...demo this. Contact me via my blogsite if you want them.
ARKHAN
This is an important point - a potential bug I thnk. I'm just re-posting to make sure that it hasn't been overlooked.
-Jamie
U412
This isn't surprising either. Try to think what's happening here. Its all to do with understanding the container hierarchy.
1. A task in the child package throws an event and it gets caught by an event handler in the child package. This is as normal.
2. The event continues to propogate up the container hierarchy. Remember that the child package itself is a container and is a child container of the TaskHost container of the Execute Package task in the parent package.
3. As the event propogates up the container hierarchy like this it eventually hits an eventhandler in the parent package.
A little bit on the container hierarchy here if anyone is interested.
-Jamie