Thanks EVERYONE for the assistance on this one. Much appreciated! Great user group
The following curl batch script consistently recreates the problem. The difference between a fast and slow replication is whether or not the full URL is used for the "from" and "to" databases cls echo off echo start from scratch c:\es\curl\curl -X DELETE http://admin:[email protected]:5984/from c:\es\curl\curl -X DELETE http://admin:[email protected]:5984/to rem c:\es\curl\curl -X DELETE http://admin:[email protected]:5984/_replicator/test?rev= return echo create from database c:\es\curl\curl -X PUT http://admin:[email protected]:5984/from echo _ echo create to database c:\es\curl\curl -X PUT http://admin:[email protected]:5984/to echo _ echo create dummy document to hold attachment c:\es\curl\curl -X PUT -d "{\"dummy\":\"slow\"}" http://admin:[email protected]:5984/from/test echo _ echo upload the problem attachment - BE SURE TO SET THE REVISION NUMBER c:\es\curl\curl -X PUT http://127.0.0.1:5984/from/test/x937?rev=1-23a888c74ba58fe20dc0d424133750bb --data-binary @slow.x937 -H "Content-Type: application/octet-stream" echo _ echo time before %TIME% echo _ rem echo FAST REPLICATION %TIME% rem c:\es\curl\curl -H "Content-Type: application/json" -X POST -d "{\"source\":\"from\",\"target\":\"to\"}" http://admin:[email protected]:5984/_replicate echo SLOW REPLICATION %TIME% c:\es\curl\curl -H "Content-Type: application/json" -X POST -d "{\"source\":\"http://127.0.0.1:5984/from\",\"target\":\"http://127.0.0.1:5984/to\"}"; http://admin:[email protected]:5984/_replicate echo time after %TIME% On Friday, January 24, 2014 2:26 PM, Paul Davis <[email protected]> wrote: The important thing to note here is that test 1 is using URLs for databases where tests 2 and 3 are using internal replication. Reading the logs it shows that the request to write the attachment took about 72 seconds. Out of curiosity could you rerun test 3 twice more with the following JSON bodies: {"from":"http://localhost:5984/from";, "to":"to"} {"from":"from", "to":"http://localhost:5984/to"} This should narrow down if there's a problem with reading attachments, writing attachments, or some weird interplay between reading and writing attachments. If it turns out that the only slow version is when both are using the "http://localhost:5984/dbname"; versions then can you run one more test: {"from":"http://127.0.0.1:5984/from";, "to":"http://localhost:5984/to"} If that's also slow then we probably have a weird problem with the multipart code (fun!). If its not slow then most likely this is a bug in the HTTP worker pool implementation (also fun!).
