SCCM 2007 : State Migration Point Bug , Does not clean up the old SMP- USMT Records in "Computer Association".
A few weeks ago I ran into some serious trouble at a customer.The state migration point at a secondary site ran full. At first site this wasn't really an issue , as we expected to go over the default values that where configured in the state migration point properties as the site contained 800 users . The pic below confirmed the settings , so we altered it to the new settings .
Problem solved we thought , nothing was thru ...
As soon as we did one extra , the same problem reoccurred. Strange ?! What the Hell ?!
After doing some futher investigation , it seemed that our older records that should have been purged after 14 days , where still into the console as shown below , with also the content still available on disk .
OUR SYMPTONS:
Consider our scenario:
* On a site server that is running Microsoft System Center Configuration
Manager 2007 Service Pack 1, you configure a state migration point role.
* On the state migration point, you capture several user states and then
restore them.
* You wait for state store deletion to occur on the state migration
point.
In this scenario, you notice that only two expired user state stores are
deleted. Additionally, the Smpmgr.log file indicates that the first user
state store deletion was successful but that the second deletion failed,
and you receive the following error message:
CheckAndDeleteSMPStores failed with error code (80004005)
Looking in the smpisapi.log , everything looked ok :
========== BEGIN: smpisapi.log ==========
Query string to parse: 'op=healthcheck'.
Creating state store - E:\SOURCES\USMT\SMPSTOREE_B2DE4AB0$\FBC02D2321401636EFBD99756C5AAF338C3D018C06B9724EB56B409163928473. Statestore - E:\SOURCES\USMT\SMPSTOREE_B2DE4AB0$\FBC02D2321401636EFBD99756C5AAF338C3D018C06B9724EB56B409163928473 exists.
Returning status "200 OK"
========== END: smpisapi.log ==========
After some further investigation in the smpmgr.log .It is really the message that he is not able to delete the stale records!
It does not perform the cleanup as scheduled into the SMP settings. We have specified 31 Days for the cleanup to happen and the folders with the content are still there !
**************************smpmgr.log***************************************
Checking store for cleanup of failed or stale state stores...
Cleanup of store "E:\SOURCES\USMT\SMPSTOREE_B2DE4AB0$\FBC02D2321401636EFBD99756C5AAF338C3D018C06B9724EB56B409163928473" failed with errorcode (80004005)
CheckAndDeleteSMPStores failed with error code (80004005)
Failed to cleanup stores. (0x00000000)
Secondly we get the Http error 500 back into the log file , so it seems that both are related;
Completed availability check on local machine
**************************smpmgr.log***************************************
What is causing this :
When the state migration point module deletes the first in a series of expired user state stores in a loop, the module does not correctly
initialize the computer-name length for the next iteration of the loop. Therefore, after the second expired user state store is deleted, you
receive the error message, and additonal user state store deletions are aborted.
What is the solution :
After having a long session with Microsoft Premier Support we came to a non public hotfix solution , but that has been publicly made available from the 26th of Januari !
To apply this hotfix, you must have System Center Configuration Manager 2007 Service Pack 1 installed.
This hotfix KB961269 does not replace any other hotfixes!
The hotfix wil replace the Smpmgr.dll with file version 4.0.6221.1147 and will do a site reset after .
Hope it helps ,
Kenny Buntinx