Thanks for the response. I should probably ask a different question.
I’ve been using git to take snapshots of my databases in certain states. Basically i put all the data files under git and then when I have my data the way I want I can freeze it for later replay by committing it to a branch in a repo. It works really well for me. I’m thinking of the similar mechanism for backup. Basically I’ll zip the data directory and store it in S3 periodically as a backup. I was concerned though that the state of the files might not be restartable if I don’t shut down the couchdb process before snapshotting. Anyone else doing something like this? > On Jan 14, 2016, at 1:00 PM, Robert Samuel Newson <[email protected]> wrote: > > [couchdb] > delayed_commits=false > > It’s false by default from 2.0 onward too. > > When set to true, a timer calls fsync once per second. I’d argue your forcing > isn’t necessary in either case. > > B. > >> On 14 Jan 2016, at 17:46, Dan Santner <[email protected]> wrote: >> >> Is there a way to force couch to flush everything to disk? right now I’m >> stopping the database to make that happen but would be really nice to be >> able to do that while keeping it running. >> >> >> Maybe this isn’t even necessary? >
