On Nov 2, 2008, at 21:19, Paul Davis wrote:
Also on second reading of the email, make sure to leverage loading documents via _bulk_docs. When I load huge numbers of documents I tend to try and load a couple thousand at a time. If an update fails I fall back to a binary search to find the offending record. There's an open request for identifying the offending records.
Yes, I do that. I'm down to 400 documents per request though - I got way too frequent failures (500 errors) when using more than that. I don't believe it's bad data, because just trying to submit it again and again will eventually work (at least that's what happens on the occasional errors when I submit just hundreds of documents).
I wrote about this a few days ago in a mail with the subject "_bulk_docs limit".
- ask
