At the moment I am supervising an AD Domain Migration with
Quest Migration Manager for Active Directory 8.7
The Directory Synchronization (DAS) with more than 400,000 objects had run stable for more than 6 weeks.
3 weeks ago we started getting messages about user accounts not having been migrated.
With the help of the Quest support I found out how to fix the problem with directory snchronization agent.
Workload for the Directory Synchronization Agent (DSA) too high
We are migrating Intra-Forest from 4 source-domains into a target-domain.
With help of Quest Support we could limit the causes and finally solve the issue. I learned quite some new tricks for troubleshooting:
The cause of the problem was the workload for the Directory Synchronization Agent (DSA) being too high.
How to notice the workload for the DSA as being too high?
After a full cycle, the DSA pauses for a while. This pause is listed in the DSA.log as follows:
„Common *** CFG: going to sleep“
Path to the DSA.log file: „Quest Software\Migration Manager\DSA\CONFIGS\DSA.log“
In normal operation we had between 30 and 50 pauses. Then, since one day 3 weeks ago, we have not had pauses anymore.
Note: the share ware tool Total Commander can search the packed Logfiles in the LogArchive for text.
The DSA processes some changes in a so-called „rerequest queue“. It can only be processed after a pause of the DSA. In our case, some newly created user-accounts got into this „rerequest queue“. This of course, let to these accounts not being migrated. As a cause we could detect a Metadirectory service where the same changes were written in a never ending spiral. These changes occupied the DSA in such an extent that there was no longer time for a pause and the „re-request queue“not being processed.
How to notice the processing of the „re-request queue“ in the Logfile?
When objects for further processing are send to the „rerequest queue“ the following entry is listed:
Target JobID:0 <- 4 objects sent to rerequest queue
The processing of the rerequest queue is recorded with the following entry (not following each other directly)
Common JobID:0 <- Retrieving next page (synchronization mode)
Source JobID:0 -> Forced Request Handler start
Source JobID:0 -> Forced Request Handler finish