| 6:14 pm on Nov 8, 2012 (gmt 0)|
So, a request for www.example.com/thispage on the old site returned 404 and now you redirect those requests to the new site?
Yes. Google will note the redirect and schedule the new URL for spidering at a later date.
Those requests will of course return 404.
The solution was to not have redirected requests that previously returned 404 on the old site.
It's too late to do anything now. Google will request the URLs on both sites forever.
There's nothing else to do other than ignore the list of 404 errors. The list is there in case you need to bring a page back that you didn't know was missing.
| 6:28 pm on Nov 8, 2012 (gmt 0)|
a 404 can't get passed if you provide a 301 or 302 status code and a 404 can't cause a redirect.
| 6:31 pm on Nov 8, 2012 (gmt 0)|
I might have worded this wrong. In the webmaster account they would list the pages we had 404 errors for. Some would be example.com/blog/widget1 even though there was no page like this. There was a page example.com/widget1. Now we are seeing something similar in the new site, but they'll be pages like example2.com/2011/widget1 even though the actual page is example2.com/widget1. Could the same bug be causing both errors and this bug or glitch is being transfered through the redirects we are doing on individual pages from 1 site to the next site.
| 6:41 pm on Nov 8, 2012 (gmt 0)|
are your log files showing any referrers for those urls?
| 7:35 pm on Nov 8, 2012 (gmt 0)|
Not that I can see.
| 12:30 am on Nov 9, 2012 (gmt 0)|
|a 404 can't get passed if you provide a 301 or 302 status code and a 404 can't cause a redirect |
A redirect will always be followed by something. The googlebot doesn't go straight there like a human, but it still puts the new name on its shopping list for later.
The OP here is doing a blanket domain-to-domain redirect. So anything that didn't exist on the old domain will similarly not exist on the new domain.
Copywriter, it sounds as if you still have the old site although you're not using it. If there is any pattern to the nonexistent filenames, intercept them before you do the global redirect. A 404 upfront is better than a 301/302 leading to a 404.
Do any of the nonexistent pages come with a referer? Sometimes it's a typo in someone else's site. Sometimes it's just g###s fevered imagination ("no data available")-- which is infuriating but beyond your control.
| 1:14 am on Nov 9, 2012 (gmt 0)|
|a 404 can't get passed if you provide a 301 or 302 status code and a 404 can't cause a redirect. |
Requests for URL A return 404 status on site A as there is no such page.
You start a new site, site B.
After a while you redirect all requests for site A to site B.
Some time later, Google starts asking for URL A on site B. This happens because URL A on site A no longer returns 404, but instead returns a redirect to site B.
| 2:17 pm on Nov 9, 2012 (gmt 0)|
But it doesn't seem as if any of the pages being shown on the second site are duplicates of anything on the first. Yet, with both sites we are getting a lot of 404 errors for fictitious pages.
| 9:52 pm on Nov 9, 2012 (gmt 0)|
Oops, backtrack. I think some of us assumed that the 404s on the new site are the end of a two-step process: blanket 301 from old to new, leading to 404 on the new.
Are the 404s from people coming to the new site "cold"? Apart from domain name, are they the same URLs that are/were getting 404s on the old site?
Most important: Are you getting human 404s, or do they come strictly from the googlebot? If it's only google, you can just ignore them. Mark them "fixed" and carry on.
| 12:38 am on Nov 10, 2012 (gmt 0)|
I think these are all from googlebot. What happens if I mark them as fixed even though technically I haven't done anything with them. Does that mean they ignore them?
| 1:55 am on Nov 10, 2012 (gmt 0)|
Sometimes they come back after a while. Generally they don't. In this context, "fixed" simply means "yes, OK, I know". If there isn't a mistake at your end, there's nothing to fix.