Environment
Novell eDirectory 8.8.5 for All Platforms
Situation
Server reports error -699 in replica sync for the partition root object in traffic between specific replicas.
Other servers report -625 errors in communication between the affected servers.
A trace showing inbound sync traffic does not show any error.
A trace showing outbound sync traffic shows the following error:
Sync - Process: Send updates to <server.context> for .partition.org failed, fatal (-699).
Other servers report -625 errors in communication between the affected servers.
A trace showing inbound sync traffic does not show any error.
A trace showing outbound sync traffic shows the following error:
Sync - Process: Send updates to <server.context> for .partition.org failed, fatal (-699).
Resolution
Looking at the problem from iMonitor, the -699 error is only shown on the sending server and points out which server is the receiving server. Enabling inbound sync trace on the receiving server, the following message can be seen:
Sync Point Time Difference: 346894, Time Out: 10800, Originating Agent: server1.context., Requesting Agent: server1.context
That is the only information regarding the synchronization attempt between the failing server and the receiving server. Looking at the partition root object of the affected partition in the receiving server, notice that the following two attributes are present:
syncPanePoint
syncWindowVector
These attributes are normally used during the synchronization process to mark how far the synchronization process has been able to synchronize, but normally they will be cleared after a completed sync cycle. If the attributes are remaining behind, then they can cause the problem described here.
In order to fix this issue, determine the Entry ID of the affected partition root object on the receiving server. Then perform a single object repair on the object to clear the syncWindowVector. This is the syntax to perform this on Linux/Unix:
ndsrepair -J <entryID> -Ad -sx "syncWindowVector"
This is an advanced and potentially destructive switch. Use with caution and if you have doubts on how to perform this process, contact Novell Technical Services to assist you in the process.
If the problem is present in multiple replicas, repeat this process until the -699 errors have been cleared.
Sync Point Time Difference: 346894, Time Out: 10800, Originating Agent: server1.context., Requesting Agent: server1.context
That is the only information regarding the synchronization attempt between the failing server and the receiving server. Looking at the partition root object of the affected partition in the receiving server, notice that the following two attributes are present:
syncPanePoint
syncWindowVector
These attributes are normally used during the synchronization process to mark how far the synchronization process has been able to synchronize, but normally they will be cleared after a completed sync cycle. If the attributes are remaining behind, then they can cause the problem described here.
In order to fix this issue, determine the Entry ID of the affected partition root object on the receiving server. Then perform a single object repair on the object to clear the syncWindowVector. This is the syntax to perform this on Linux/Unix:
ndsrepair -J <entryID> -Ad -sx "syncWindowVector"
This is an advanced and potentially destructive switch. Use with caution and if you have doubts on how to perform this process, contact Novell Technical Services to assist you in the process.
If the problem is present in multiple replicas, repeat this process until the -699 errors have been cleared.