Multiple Sync Gateway nodes incrementing _sync:seq is routine behaviour - it shouldn’t cause any issues like the one you’re describing. Sync Gateway 2.1.1 should be releasing any unused sequence numbers, so I’d like to better understand the scenario where a sequence is allocated and then doesn’t get used until step 5.
Some questions to help understand what you’re seeing:
- In your step 2, how many documents are you pushing to each SG node? How are you writing the documents (PUT doc, bulk_docs, etc). Are the writes successful? What are the sequences associated with the new revisions?
- What is the skipped sequence of the ‘skip’ in step 3?
Thanks.