I had about a million pages of a catalog indexed under a subdirectory, and now that's moved to a subdomain. GoogleBot is crawling each one of them and getting a 301 redirect to the new location. Even though I have set up the redirect rule in the apache sites-enabled configuration file, (i.e. it's early on when apache does the redirect - PHP is not even getting loaded), even though I have done that, the server isn't handling the load well. GoogleBot is making around 5 requests per second, and on top of my normal traffic that is hiking up the CPU for a few hours at a time.

I checked in Webmaster Tools and the corresponding documentation for a way to let Google know that the content had been moved from a subdirectory to a subdomain, but with little luck. Basically the most helpful thing I saw said to just send 301 headers for the new location.

How can I tell GoogleBot that a subdirectory is now a subdomain? If that is not an option, how can I more efficiently send 301 redirects out for a particular subdomain?

I was thinking perhaps the Nginx server but I'm not sure that I can run both Apache and Nginx side by side on port 80 for different subdomains.


2 Answers 2


You mention that you already read the documentation and you are using the right steps, so then that's done. Of course you can post here the specific technics you are using and then we can check that you are doing it correctly.

About the redirection, 301 is the best option. If you moved something permanently, then 301 is what you have to send to the clients.

You also mention the speed that Google is crawling your site, then you should control that, check this, Changing Google's crawl rate, plus remember that you can control the frequency of revisits in the sitemap. That won't help for the first visit, but as soon as it finishes wit it, your problem should go away and with the right schedule in your sitemap, everything should be ok.

You can also block parts of the site using your robots.txt. Google respects that. Then, after a few hours, you can edit the robots.txt and allow access to some other folders. That will control flow to your sites, but may cause delays in indexing and appearance in Google results. It's your decision.

Don't forget to add the subdomain to your webmaster tools account so you can control all the settings, including the crawling rate and frequency.

Check your logs to see that all the redirections are working as expected.

About having two web servers listening in the same port, you shouldn't do that, that will make them collide and break. Plus I don't think you can do that with modern servers since they check the port availability before starting.

  • Thanks, I think that these are generally good suggestions. Nothing here that I'm not already doing, except I don't think that blocking the URLs with robots.txt is a good idea because that is sending the wrong message to GoogleBot - if anything I should probably send something like an Error 503 (temporarily unavailable). Logs confirm the 301s are working. I am going to look into using maybe a reverse proxy or something to run both webservers...
    – cwd
    Oct 28, 2012 at 16:53
  • a 503 is a good option indeed. If you do that, remember to add the Retry-After header and set it with enough time, for instance, assign each main folder/directory a margin of 6 hours (just an example), o the delays will be 6, 12, 18, ... hours.
    – PatomaS
    Oct 29, 2012 at 0:14
  • Also, you should talk with the hosting company, they can see more logs and check what is causing the stop in the connection and the drop of data. It doesn't seem to be the number of clients allowed. The timeout some will say that is too long and should be 1 second, but I think that's what you have is ok.
    – PatomaS
    Oct 29, 2012 at 1:01
  • I would like some other answers on how to do this but I'm going to go ahead and accept this one even though it basically just recaps best practices and doesn't give me anything break-through. Thanks.
    – cwd
    Oct 29, 2012 at 3:01

GoogleBot is making around 5 requests per second, and on top of my normal traffic that is hiking up the CPU for a few hours at a time.

You can reduce the rate google and other search engines are crawling with the robots.txt "Crawl Delay" directive. For example, set the maximum until things cool down: Crawl-Delay: 30

I agree with everything PatomaS said.

  • I read that GoogleBot does not pay attention to Crawl-Delay in robots.txt and that it needs to be set in the Webmaster tools.
    – cwd
    Oct 28, 2012 at 16:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.