Connector Space Object don't match Full Import Object Count

Just want to clear up something

I have an HR system that removed about a 100 users. My anchor attribute is Employee_Person_ID (e.g. 12345). My join rule is EmployeeID (e.g. 123456789) 

So HR removed the records and if I do a Full Import I see the correct amount of objects e.g. 1000. I confirm this my doing a database query and I cannot find the delete objects. If I do a search on the connected space objects I get 1100 user records.

Problem 1: I still see the old objects in the connected space and they still have a connector to the MV object.

Problem 2: 

HR re-added them with new records. The EmployeeID stays the same (e.g. 123456789) but the Employee_Person_ID changed to (e.g. 54321).

Now I get ambiguous-import-flow-from-multiple-connectors errors caused by problem 1.


May 14th, 2013 9:25am

To solve your more immediate issue, you can clear the connector space and re-import all the records.

I've long suspected that the obsoletion on the SQL MA doesn't work properly, though not sure if it's even been raised as an actual issue.

Essentially, when FIM performs a "Full Import", it's meant to detect that certain records in the connector space are no longer present, and report those as deletions. But in my experience, there seem to be cases where this doesn't happen.

Anyone else want to weigh in on this?

- Ross

Free Windows Admin Tool Kit Click here and download it now
May 16th, 2013 2:23am

Hi Ross,

We've had the same experience with IBM DB2 MA's in past so it's not entirely SQL related to the SQL MA...

Regards

Jacques.

May 16th, 2013 5:34am

That doesn't surprise me either Jacques. I believe I've seen it in other MAs.

I guess the question is - is this something that is known by MS?

- Ross

Free Windows Admin Tool Kit Click here and download it now
May 16th, 2013 7:26am

Hi 

Yes I was also expecting to see the deletions after a full import. I know this works as I have see this before. First time I see this not happening.

This is an Oracle MA and not sure if this is the issue. 

The environment is running on FIM 2010 Update 2. (4.0.3606.2)

May 16th, 2013 7:49am

I would use the connector space delete option as a last resort. This is not a uncommon scenario and will happen from time to time. Would like to find a resolution. 
Free Windows Admin Tool Kit Click here and download it now
May 16th, 2013 7:52am

"Problem 1" is not really a problem but by design.
If you look at the properties of the affected objects, you will notice a staged deletion.
This is necessary to enable sync to detect deletions.

For more details on this, see Understanding Deletions in ILM.

"Problem 2" is a result of  "Problem 1" - sync is state based.
You have now a staged add and a staged delete for the same object - now what? :-)

The solution is very simple - just run a sync on your HR MA to get rid of the objects with a staged deletion.

Cheers,
Markus

May 16th, 2013 12:18pm

Hi Markus,

Problem 1. 

I also expected an stage deletion. But if I look at the object in the MV it still shows as an connector. If I do a direct Oracle query the object does not exist in the DB. If I do a CS search the object still exist there too.

Problem 2

  I have done multiple full imports and full sync's on the oracle MA and does not show any deletes on the MA.

Free Windows Admin Tool Kit Click here and download it now
May 16th, 2013 12:58pm

Help me out - I'm missing something...

The object must show up as connector (after an import) - this is OK:

  • When you run a sync, the object is removed.
  • When the object is removed, the link to the metaverse object is removed.
  • The removal of the link triggers the object deletion rule on the metaverse object.

So, this is all - as far as I can tell - OK.

Just making sure, are you saying that the "deleted object" still shows up as connector AFTER running a sync?

Cheers,
Markus

May 16th, 2013 2:04pm

I agree with your statements and this is what I also seen on other environments. 


The object must show up as connector (after an import) - Yes, there is a connector

  • When you run a sync, the object is removed. No, this is not happening. CS object count does not go down. No deletion shown.
  • When the object is removed, the link to the metaverse object is removed. No this is not happening. Object still present in the CS and has a connector in the MV object.
  • The removal of the link triggers the object deletion rule on the metaverse object. No, I cannot see this happening


Just making sure, are you saying that the "deleted object" still shows up as connector AFTER running a sync? Yes!

Summary

RecordNo EmployeeNo Details

ABC 1234567 Original Record with a connector to the MV object. This records gets deleted and confirm with direct DB query. FIFS does not show a object delete on the MA. CS search still shows object.

DEF 1234567 New record arrives with ambiguous-import-flow-from-multiple-connectors error. The only why I can get rid of the error for now is to make the old record a explicit disconnector. Even after this the old record still shows up in the CS search.

Free Windows Admin Tool Kit Click here and download it now
May 16th, 2013 2:38pm

Ahh - OK.
Have you tried running a preview on an object that should have a deletion staged from an import only operation?
The combined run profiles (FIFS) will be removed at some point and it is advisible to stop using them.

So, one question is whether you are dealing with an import related issue.
If staging doesn't work as expected, sync must fail...

Also, just making sure, you are not dealing by chance with re-provisioned objects - right?

If the affected objects don't have a staged deletion after a FI (only), you need to open a case with CSS.

Cheers,
Markus

May 16th, 2013 3:41pm

Thanks for the replies Markus,

I did FI and the count under statistics (9475 + 2 duplicate object = 9477) match the DB query of 9477. But if I do a CS search I get a total of 9580. A 103 objects more. I have made them explicit disconnectors for now. But I expected them to to be deleted out of the MV as they do not exist in the HR system anymore, or am I missing something here?

Thanks for the Tips on FIFS.

Free Windows Admin Tool Kit Click here and download it now
May 16th, 2013 7:36pm

Markus - the important thing to note here is that the obsoletion fails and the deletions do not trigger,  so there are no staged deletions set on the CS object, so no amount of syncing will fix this, as FIM believes the objects are still in the source.

I may also be wrong here, but I believe if the object was flagged for deletion, that they wouldn't be seeing the ambiguous import error.

I'll see if I can reproduce it, put together a package and log it as an issue on connect (That's the place to log bugs, right?)

- Ross Currie

May 17th, 2013 1:12am

Thanks, Ross.

Connect is only intended for bugs that are related to connectors that don't ship with FIM.
All other cases need to be handled by CSS.

Cheers,
Markus

Free Windows Admin Tool Kit Click here and download it now
May 21st, 2013 3:21pm

Connect is only intended for bugs that are related to connectors that don't ship with FIM.

All other cases need to be handled by CSS.

CSS is Customer Service & Support, right?

In this case, I've discovered a bug and want to report it to the product team. I can work around it, so don't need CSS to resolve it for me, but it makes sense that the product team be notified somehow.

How do I go about that? CSS does not appear the logical choice here.

- Ross Currie

May 31st, 2013 7:08am

Hi Ross and mybm3w, just curious if you ever found a fix that didn't require the deletion and re-import of the connector space? I too have a customer with an Oracle MA with 29,000 objects in the source system but 41,000 objects in the connector space. Cheers. David
Free Windows Admin Tool Kit Click here and download it now
March 23rd, 2015 9:58pm

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics