Provisioning Dispatcher error: Clear ALL old Semas

[Former Member]

So I've gotten this periodically in the past, but recently we've been getting randomly spammed by this error message in our System Log:

"Clear ALL old Semas:  Exception:java.lang.NullPointerException"

It has no additional details, Job ID of 0, and seems to be coming from our second provisioning dispatcher, but I'm not 100% positive they haven't both done it at some point. Is this related to having two Provisioning dispatchers running on the same machine? That's my only guess, as I've heard people mention that being a problem in the past, but never exactly WHY it is a problem.

Alternatively, is there a way I can find out what code it is actually running so I can fix the exception handling or maybe add a check or two to avoid trying to use a null object (in the end, I'd probably consider this a code fault regardless of what's causing it).

Thanks!

[Former Member]

Hi,

There is a checkbox for automatic startup of dispatcher under Managemant\Dispatchers\<name of the dispatcher>

also check the value for check frequency and reload frequency.

[former_member2987]

Adam,

There's two things I can think of off of the top of my head:

1. Are you using Audit Flags? Also what's in your MC_SEMAPHORE table?

2. Do you have any lingering jobs in the provisioning queue?

Matt

[Former Member]

Yes, I am still getting this error.

M Pollicove:

Yes to both of your questions.

There are a few places that we are using audit flags, though I think most of them aren't actually important or are deprecated and could be removed, if that is related to the issue. I wouldn't have guessed it would have anything to do with audit flags, but stranger things have been true before.

There are definitely a LOT of random jobs sitting around in the provisioning queue, quite a few of which I assume are dead / stalled / orphaned at this point and could be killed, it's just a bit scary to go around clearing out the provisioning queue in a production system. I'll see if I can knock out a few of them, but there's no way I can clear all of the dead ones easily.

[former_member2987]

Adam,

I'll be honest with you, I'm guessing on both of these but here's my thinking:

1. Clearing the provisioning job of dead/stale jobs is a good idea. I'll have to do some reasearch to see if there's a best practices on this. While it might not necessarily fix the current issue, it would not hurt.

2. The original name of the audit flag was semaphores. So there's your connection.

What might be related to #1 is Ian Daniel's Blog link: http://sap-idm.blogspot.com/2011/01/useful-sql-for-sap-idm-for-provisioning.html

Might give some visibility.

Anyone have any best practices on clearing the provisioning queue? Was a lot better with the monitoriing / audit interface in 7.0 / MaXware Identity Center. I think they're working on getting some of that functionality back... We'll just have to wait and see.

M

[Former Member]

I'm working on cleaning up the provisioning queue now, starting with entries that are more than 60 days old, as the longest approval we allow is 24 days so anything that old should be guaranteed dead.

Also, I forgot to mention this in my last post, but the MC_SEMAPHORE table is empty.

Thanks for your advice!

[Former Member]

Here's a new one:

Clear ALL old Semas:  Exception:com.microsoft.sqlserver.jdbc.SQLServerException: The native type 0 is unknown.

This one is from our regular provisioner... Now we're getting both this AND the other one.

[Former Member]

I thought the problem might be related to incorrect configuration of our dispatchers (we had two provisioning dispatchers and two regular dispatchers running on the same machine apparently), so I fixed the configuration of those and thought the problem went away.

This morning, it has come back. It started with exactly one of this message:

Clear ALL old Semas:  Exception:com.microsoft.sqlserver.jdbc.SQLServerException: The native type 4 is unknown.

Then over the next two hours we got this message in our log over 1600 times:

Clear ALL old Semas:  Exception:java.lang.ArrayIndexOutOfBoundsException: 3

Does anyone else have any thoughts!? This problem is plaguing us right now... I'm not actually sure what it is doing that is bad, but it's filling up our log with junk and we're trying to measure how well our system is working using error counts, which this shoots through the roof.

[former_member192837]

We also got same kind of issue few days back.

provisioning dispatcher was throwing error "Clear ALL old Semas: Exception:java.lang.NullPointerException".

The error does not harm the system functionality and is rather a warning.

we checked the system and found there was no issue with fetching the data , and not effecting provisioning as well .

It is just filling the system log unnecessarily. Restarting the concerned dispatcher solved the issue.

After a restart the error did not occur anymore.

[Former Member]

We have also seen this issue and it keeps re-occuring. As a sort of quick fix we have set up a litlle service on the DB that checks for this state -and if it occurs, it stops and starts the dispatchers.

This way we do not need to manually stop and start dispatcher.

Kind regards

Heidi Kronvold

[Former Member]

What type of Database are you using? Is it MSSQL? Can you provide details on exactly how you set it up to detect and auto-restart the dispatchers?

[Former Member]

Hello Adam

Did you find a way to fix the error... We have set our log on the production system to 10000 entries, but 8000 of these seem to be "Clear ALL old Semas: Exception:java.lang.NullPointerException: while trying to invoke the method com.microsoft.sqlserver.jdbc.TDSReader.reset(com.microsoft.sqlserver.jdbc.TDSReaderMark) of an object loaded from local variable '<8>'"

At the moment, I don't even know where to start looking...

Best regards,

Anders

[Former Member]

Hi..

Have you got answer or solution for this error....we are getting error as below

The dispatcher "mq_job" was constantly throwing error "Clear ALL old Semas: Exception:java.lang.NullPointerException".

Date;Type;"Dispatcher";Job ID;"Additional info";"Text"

2011.02.17 05:46:55;Error;"mq_job";0;"";"Clear ALL old Semas: Exception:java.lang.NullPointerException"

2011.02.17 05:46:50;Error;"mq_job";0;"";"Clear ALL old Semas: Exception:java.lang.NullPointerException"

Please show some way to indentify the error....

Thanks,

Kishore