Scary 2 : links disappeared on Google Site search
A bit of shock to me today, only two links to this site showed up when I did a Google site search on this site. Two days ago, there was over forty! What a coincidence, yesterday I just read on inphotos.org about such a scare. An offending ‘&’ as category name caused the sitemap by UTWgoogleSitemaps considered to be corrrupted by Google. This site’s sitemap is automatically generated too, though by a different wordpress sitemap plugin by Arne Brachhold.
I manually eyeballed the sitemap.xml and found no ‘&’ in sight. A recent post on how to minimally upgrade wordpress.org from 2.0.6 to 2.0.7 has plenty of ‘&&’, a ‘logical AND’ in UNIX/Linux shell programs. I can’t think of anything else. Puzzled, I went to Google’s webmaster tool and check out diagnostics. The sitemap is checked green and forty links are noted. The count of forty agrees with my manual count using sitemap on the server.
The only things that stand out on the diagnostic page are two 404 errors (pasted below). I know that require and require-once are PHP functions used in various wordpress.org blog server programs, and they definitely should not have exposed literally through generated HTML pages. There’s no such URL defined in the sitemap either. Therefore I am a bit at loss where Google bots got the idea there are such links on this site.
| URL | Detail | Last Calculated |
| http://www.supportsmb.com/function.require | 404 (Not found) [?] | Jan 18, 2007 |
| http://www.supportsmb.com/function.require-once | 404 (Not found) [?] | Jan 18, 2007 |
The sitemap on Google was verified on Jan 18 and it was updated on the server yesterday. To be sure, I went ahead to submit it again. The updated version came through green as well. Now I am really puzzled.
I guess I’ll give it a few days and see whether it is a temporary glitch. As some pointed out, this could happen since because things may not get replicated to a new Google data center right away, and it happened to have answered your site search queries.
Fingers crossed…
experts8 said,
January 21, 2007 @ 4:18 pm
After a afternoon nap, now it has only 1 link left*. The cached home page to the site. The cache was from Jan 17, the last day the site was under Connections theme. Since then the site moved to almost-springs for good.
Technically there were 13 links. However, if 12 of them were for sysmon.supportsmb.com. The latter was a dedicated virtual host we had, before it was decided on Jan 18, that too much repeated work to maintain one virtual host per topic.
experts8 said,
January 21, 2007 @ 4:46 pm
just upgraded sitemap plugin to 3.0beba-b5. It has more features (memory/time limit, custom XSTL) and support Yahoo! as well. You can find it here. Manually regenerated the sitemap to be sure.
experts8 said,
January 22, 2007 @ 11:57 am
This morning the old site total disappeared for good. Only two links showed up for Google site search. The home page stays, while a post from yesterday, ahem, this post, is the only other link.
Meanwhile, if you do a regular Google search by double quoting some phases unique to this site, you’ll find the page or post on this site properly.
This means Google knows more than it knows it knows. Try “site:www.supportsmb.com” then “how to secure wordpress”, you’ll see what I mean.
experts8 said,
January 25, 2007 @ 11:00 am
Finally it is all back as of this morning (~50). wasn’t yesterday (~3) from site search. Google search always showed it indexed more than what site search showed.