Skip to content

Low Threads No Matter What I Do

Hey Guys, 

I bought GSA SER and have been meaning to restart my SEO efforts for a while. Finally cleared out time to get it done. 

A lot of people are concerned with LPM...my concern is more basic, the active threads. 

I have GSA SER running on a fairly powerful VPS, 3.6 ghz quad core, 4GB or RAM and a 1gbps connection. Speedtest.net shows around 300mbps but that's normal cause it's not realy meant to test gigabit connections. I have SER configured to use 300 threads with a html timeout of 60 (previously was 30 but I tried changing it to see if it would perhaps somehow help with my active threads, despite my better judgement). I use 100 semi-dedicated proxies from buyproxies and both the GSA and Scrapebox show the proxies as working, though SER says that all 100 are functioning while Scrapebox says that around 70 are working but that shouldn't matter that much. I also manually tried using a few and they are working flawlessly. 

I use the private proxies for: Search Engines, Submission, PR check and verification and have the custom wait time between SE quires set to 10 seconds. 

For engines I have 4 loaded, Google international, google blog, Yahoo and MSN. Though I tried adding a lot of different ones like multiple Google, Yahoo and MSN countries as well as the goduck search engines but to no avail. 

I've read up on here about similair issues and I hoped that my low keyword count of only a few hundred might be the culprit. I scraped 20k keywords with Scrapebox and imported it into the keyword list for that project and that seems to have helped a little bit. Instead of just hovering around 5-10 threads now it runs up to 50ish from time to time but it always drops down again to 5-10. 

I don't really know what else I can do to improve this, if anyone has any suggestions they're greatly appreciated! Thanks in advance 


Comments

  • Something I forgot to add, CPU usage never goes over 20% so it probably isn't an issue of processing power. RAM is at around 1GB (of 4 available)
  • How many projects ?
  • edited July 2013
    This probably isn't relevant but I only used 1 email account for this project, hotmail. 
  • LeeGLeeG Eating your first bourne
    edited July 2013

    Anything over 100mbt connection is wasted on ser in all honesty

    Im running 200 threads and averaging about 10mbt, 20mbt peaks

    HTML timeout increase to about the 130 to 140 area

    Then keep an eye on your verified. Too many failed, increase the time

    Things that affect LpM / thread count

    proxies

    search engine choice. Ozz shared a method of using random engines on here

    https://forum.gsa-online.de/discussion/comment/30966/#Comment_30966

    keywords, bigger the list the better

    engine choice, monitor the submission stats. You will see some engines produce poor submission levels. Try and target the engines that give a good submission to verified ratio

  • Leeg thanks for the great tips. 

    Already set the HTML timeout to 130

    One thing that I don't get is that SER tells me that all the proxies are working, even when checked against Google/Bing, Scrapebox tells me that 70% of them are working but I still get "ip/proxy block on Google" in the log from time to time. If SER says that they're all working how can they be blocked ? 

    I mean I understand how they can be but does that mean that the SER proxy checker is faulty or what ?

    Currently my LPM is around 4...a long way, away from your 100-250 numbers. 

    My proxies all have a response time from 0.3 to 0.7 and they should be working according to SER and Scrapebox so I'll rule that out for now

    I'll try the random search engine trick, thanks to both you for letting me know and Ozz for coming up with it. 

    Keywords are interesting, when I switched from a few hundred to 20k it did make things go faster, instead of the 0-10 threads I got up until that point now I have a steady >10, sometimes it springs up to 50s and drops again. But I find it hard to believe that it burned through 20k of keywords already so I don't really understand how a new batch of keywords will help speed this along. Would you mind explaining this to me ?


  • LeeGLeeG Eating your first bourne

    Use the Ozz method for random engines

    Chances are your using shared proxies

    Who are you sharing them with, what software are they using and what engines are they targeting. Something you have no idea about

    So chances are, it might not even be you to blame

    By using the random engines, your pulling results from different googles

    I shared a method on here once for proving my theory correct, that most googles will return similar results. At the time it was about an 8 place spread on all the googles

    I don't get many ip blocks, because I tend to steer clear of the .com google etc which most people will target for results

  • if you create your SE mods than duplicate and rename those mods without the operators (inurl:, site:, ...). if you getting troubled with IP/Proxy blocks than you can try to use this 'non-operator mod'. i believe that google won't block you so fast whith this.
  • The IP/Proxy Blocked by Google is not that common that it's my main concern. It seems to happen on queries like ""Powered by Blogtronix" "Attached Image: "" 

    My main thing for now is to get the threads up there into the hundredths. 

    I don't think it's an proxy issue but it's certainly possible. 

    I'll try what you suggested as well as getting the keyword list into the 100k and see how that does it. 

    Thanks LeeG and Ozz, your help is much appreciated! 
  • Interestingly enough...using a few of the proxies I use in GSA manually gives me this: 

    image
  • OzzOzz
    edited July 2013
    click to view page #10 or change settings to 100 results per page.
  • Interesting, only 10 pages displayed for this search. 

    Why do you suppose that is ? Is it cause Google doesn't allow more for this specific search ? 

    I tried searching other quoted keywords and it shows them just fine beyond #10. 

    What does this mean for me ? 
  • Try add to your query "&filter=0" and you will see more results.
  • Here's an update. 

    I used Ozz method for the search engines and I've created 5 just like he instructed. I also scraped a list of some 50k keywords and got my thread count up to around a 100 and had an initial LPM of 19-21 which was fantastic for no!

    Unfortunately, about  5 minutes in the thread count dropped to the 20s and a few minutes after that to under 10. 

    I think I understand the problem now though. It's all about the keywords. 

    The first time around, before I opened this thread I would input like 5 keywords into the Scrapebox keyword scraper and set it to level 4. I guess the keywords were just too similar to get a good thread count out of them, for whatever reason. 

    This time around I used a dictionary list and level 2 scrapping in scrapebox and results are somewhat better, at peak times I get 50-60 active threads and otherwise it hovers 20-30. 

    The next thing I'm gonna do is get the dictionary list, and not go any level deep. The diversity of the keywords should significantly improve my thread count and in doing so my LPM as well. 

    I'll post an update once I've tried this. 
  • Nope.

    I tried scraping the keywords as described above and I consistently only get 30ish active threads, sometimes it goes to 60, 100 but always drops to around 30.

    What I don't understand, and maybe someone could clarify this for me is why do the keywords have an effect on the active threads ?

    I'm not sure what else I can do at this point, I just can't consistently get the active threads above 100. Any suggestions in this regard would be highly appreciated. 
  • When SER scrape search engines, threads are low. And when he submits threads are high.
    So if you have bad keywords or not enough kws, your log is full of 000/000 and SER will continually try to scrape with no success.
  • I just watched the log for a few minutes, the 000 you speak of happen every 20ish seconds once. So it's all normal operation and then it happens once and then all normal and then again. I obviously don't know for sure but that doesn't seem to be something that's happening that much so I'm not sure if that's the problem. 

    On a different note, when I stopped the project and started it again I had maxed out threads for a few minutes, when I had the max at 300 it was 300, at 500 it was 500 again, then it dropped to around 200 after a few minutes then to my usual 50ish after that. 

    Any advice would be really appreciated and thanks for all the amazing help so far!
  • I usually use global list and feed SER with imported lists, and my threads is always at maximum.
    But i noticed if i disable global list and let SER scrape, threads are lower when it verify emails and scrape search engines.
    Try to run 10 projects together this will greatly increase threads count.
  • OK I already guessed that running multiple projects will do that but what if I want to build a lot of links towards one site for example without creating multiple projects ? 


  • i had similar problems, this is what i did.
    i used movies subtitle files from 19 languages, brake them down to 39 million words (remove duplicates), split them in files  1000 words per file add them to SER %spinfolder%, using around 60 search engines,
    Start with select all search engines keep your log file, spend 10-15 minutes daily to see which search engines are not compatible with what you are doing and remove them.

    also you have to modify your engine files, remove operators and all other spare decorations.

    These will keep your threads  busy full time.......

    ..... then you have to start worrying about your low LPM 
    :D
Sign In or Register to comment.