Dedi Servers Internet connection: 1Gbit/s
Url Harvester: depends on engine and proxies.
Yahoo -10 connections with no proxies
Google and AOL - Use private proxies at 20% of my proxies. So if I use 100 private proxies I set connections to 20. Can harvest for quite a while like this, and private proxies are so fast, I find my urls/s come out more then more public proxies, plus IMHO you don't have to mess around with failed ones, as 99.9% of them succeed first time with private proxies
Fast Poster: 500 - 60 sec timeout
Link Checker: 500 - 90 sec timeout on small runs, lower connections seem to be faster on my server when the list goes over 400K
Proxy Harvester: 100 - 15 sec timeout.
~~~~~~
When I used 100Mbit dedi servers
Link checker 100 - 90 sec timeout
Fast Poster - 200 - 50 sec timeout
Proxy Harvester - 100 - 15 sec timeout
Url Harvester - 75 per engine, I only ran 1 engine per scrapebox instance.
~~~~~~
Home internet - 18Mbit
Link checker 100 - 90 sec timeout
Fast Poster - 100 - 50 sec timeout
Proxy Harvester - 100 - 15 sec timeout
Url Harvester - 75 per engine, I only ran 1 engine per scrapebox instance.
I didn't see much difference in many settings as far as optimal from my home 18Mbit connection to my server 100Mbit except that the server just did everything faster. When I jacked connections success seemed to fall.
When I upgraded to my 1 Gigabit I am able to max everything and it sails smooth as can be.
What I do is start low, and turn it up until speed and results tend to fall off rather then increase and thats about where you need to be. Of course its subjective, cause the larger the file you work with, the more the processor has to compensate with, even if there is enough RAM, and of course if there is not enough RAM then you have issues.
Was wondering this and been trying optimize for last 8 hours lol.
100mbs
Harvester 350 per Engine works fine
But the link checker hangs no matter how less threads i put in . Then it goes really fast for little time and slows down again. Without proxies. This is killing me as im trying to think how to get it working better. Win server 2003.
Fast Poster 100 threads - 90s timeout
This happens to me even on my 1Gbit server when checking links at 500 connections. So long as your not getting lots of timeout errors then its fine. Basically what happens is you have a lot going on there. Windows uses a message Que to tell the scrapebox window to repaint (refresh what your viewing) and when you have tons happening your pc may not have enough resources to do everything Windows and scrapebox and whatever else you have running, are asking it to do. So it can't process those messages fast enough and the screen "hangs" or ghosts or even says not responding. Even my 6 core servers with 8GB of ram on 1Gbit lines do this and sometimes even for 10 mins at a go. I have learned don't touch it, the pc is doing what it needs, let it be. (or buy a faster pc for instant gratification lol - I spend too much as it is).
If you are getting lots of timeout errors, then your inet connection pipe isn't big enough to handle it and it times out before it has a chance to load the page. When you have 100 pages that all time out at a time it will seem "jerky" in motion as it skips. In which case just turn your connections down/timeouts up.
So basically not a lot of timeout errors, leave it be. Lots of timeout errors, then turn your connections down and timeouts up.
Oh, and no you don't need proxies when link checking, although its good to use the randomize setting so you aren't link checking 300 pages on 1 domain at a time.