However when querying the different data centers cache servers especially with the site: command for all kinds of sites you get crazy results.

I’ve seen cache: working, but site: not,
seen info: working, but cache: not
seen inurl: working, but site: not and vice versa

First I thought there’s a major reindexing going on,
but then I compared those site: command results with
the actualy cache dates returned for any site I just came across (it’s a small unimportant PR3 blog)

and there’s no common pattern when comparing cache dates to the site index count as you can see there are low counts for Dec 6 (4 pages to 147 pages) as well as high counts (950) … same for Dec 8 going from 538 to 1060

Hmmm – who cares about a PR3 blog I hear you say…

doing the same for database vendor oracle.com brings results from 987 pages to 194,000 pages, which is a huge difference as well

Those datacenters showing especially low counts seem to be affect by what was refered to recently as bad data push in the rollout of new data or binary versions of Google, which seems to have taken place yesterday Sunday Dec 10

For example these 2 DC are showing those low counts on several checks I made

64.233.179.19
72.14.219.99

on the other hand, a site that was recently launched is indexed there fully, having all the pages in the site command, while other datacenters do not…

are these two the two leading DCs ???

Does it matter? NO
Time to freak out? NO
Anything else to do? NO – get a cup of coffee and relax.

Let’s just wait for some days and
I’m sure Google will settle down, just as usual.

Update with some links

Dec 7 sudden drop in rankings
from 2 to 102nd

actually Matt Cutts confirmed “nothing special” in the WMW thread above