Investigate if Yioop having its own DNS cache independent of cURL speeds up crawling
administrator2012-02-10 08:19
Right now comcast connection is b/w 22-30 Mbps. Presumably,
the average page downloaded is less than the 50k threshold
Yioop imposes. This will probably be confirmed or refuted
after the following is implemented:
http://www.seekquarry.com/mantis/view.php?id=79
Also, it looks like queue_servers are being under utilized.
This suggests that the fetchers are not crawling as fast as
they could and DNS caching might be an issue. Should investigate
if this is true, and if so, implement a separate DNS caching
scheme than the one that comes with cURL.
Right now comcast connection is b/w 22-30 Mbps. Presumably, the average page downloaded is less than the 50k threshold Yioop imposes. This will probably be confirmed or refuted after the following is implemented: http://www.seekquarry.com/mantis/view.php?id=79 Also, it looks like queue_servers are being under utilized. This suggests that the fetchers are not crawling as fast as they could and DNS caching might be an issue. Should investigate if this is true, and if so, implement a separate DNS caching scheme than the one that comes with cURL.
Fixed with 097e53b