Has anyone developed a fix for this? I would like to try Eseosa's suggestion in the above thread:
"I find running transaction in batches help alleviate issue.
Using $xcount as counter - if X records, commit transaction and restart new transaction.
So far, no deadlocks detected on system."
If anyone else has gone ahead with this, it would be great to know:
1. How large are the batches in which transactions are committed (how many users should be processed per commit).
2. Which stage of the sync job is causing the issue. I haven't looked into this yet - I will need to find out which of the two transactions needs to be managed in batches in this way.