We've seen errors while using iManager to update attribute:

(Error -659) Time synchronization services has detected a problem with an operation attempting to modify replica information

My edir-phu is weak.
Initial googling (and the nature of the error) pointed to NTP issues.
I've cleaned up a few minor issues, but even from the beginning
ndsrepair -T
showed only a 1 second divergence (and "Yes" for time is in sync column). Now all servers report 0. The problem persists.

Some details
1) Not all imanager instances in the replica ring exhibit this problem
2) On a related note, LDAP updates seem to be going fine
3) We have 9 ring members (I know, too many)
4) ndsrepair -E output looks good (as near as I can tell) with all partitions reporting being up to date

I've run ndsrepair -R against all (or most all) of the replica members.

I suspect the next step is to run ndstrace on one of the servers that seems to have issues and attempt an attribute update.

But I'll be rapidly approaching the limits of my knowledge, and would love some more expert advice :-)