redgeoff commented on issue #470: Scheduling Replicator URL: https://github.com/apache/couchdb/pull/470#issuecomment-297231810 Background: https://issues.apache.org/jira/browse/COUCHDB-3391 Summary: I tested the max_jobs config and it doesn't appear to be working as I would expect. Sorry, if I missed a detail. Steps to reproduce: 1. Pull and build https://github.com/apache/couchdb/pull/470 2. Clone https://github.com/redgeoff/couchdb2-replicator-not-scalable-vagrant. I used vagrant to set up my test, but you should be able to run the tests outside a VM as well. 3. Modify [couchdb2-replicator-not-scalable-vagrant/index.js](https://github.com/redgeoff/couchdb2-replicator-not-scalable-vagrant/blob/master/index.js) and replace localhost with the correct dev DB host 4. Set the config replicator.max_jobs=10 5. Run `npm run demo` (this just runs index.js) => this results in `Unhandled rejection Error: internal_server_error` errors. If you were to modify https://github.com/redgeoff/couchdb2-replicator-not-scalable-vagrant/blob/master/index.js#L10 and use a smaller number, e.g. 10 then there would be no issue as you would not exhaust the file descriptors and db connections. My assumption is that the `max_jobs` config should throttle the replications and prevent the DB from throwing errors. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
With regards, Apache Git Services
