Have some evidence that 2.1 did not fix, still having intermittent issues even with 2.1.
Recreating replication fixes temporarily.
Getting similar child_died errors on XDCR side - but on _revs_diff:
[xdcr:error,2015-06-10T20:44:51.623,ns_1@127.0.0.1:<0.13072.75>:xdc_vbucket_rep:terminate:487]Replication (CAPI mode)2f1ae96b8df73f9fab3378668481243a/TestClient_userCall/tethr_call(TestClient_userCall/721->http://*****@172.16.112.128:9091/tethr_call%2F721) failed: {child_died, {'EXIT',<0.13057.75>, {http_request_failed,"POST", "http://Administrator:*****@172.16.112.128:9091/tethr_call%2f721/_revs_diff", {error,{code,503}}}}}
Getting this error 1000 times a second on elastic side:
[2015-06-12 02:09:38,755][ERROR][transport.couchbase.capi ] [tethr] Too many concurrent requests. _bulk_docs requests: 0, _revs_diff requests: 1024, Max configured: 1024
But I have a server thats been running it for a while to and its been working fine.