Reversion to OLDER cached snapshot (mid-month)

A "normal" database-update complexity glitch -- or what?

[Meta_Vision]

The FAQ addresses the problem of a
"stuck" very old cache page -- but what
I'm noticing is an update of the result listing
during "the dance" ... THEN, A WEEK OR TWO
LATER a reversion to an older cached snapshot.

NOTE: I saw this happen last month as well.
(At the next dance, the newer data was evident.)

WHILE I CAN UNDERSTAND that
database synchronization complexity
in a distributed computing system might
create such "temporary anomalies" --
nevertheless, I tend to expect Google's
masterful tech team to "ride the chaos
bronco" perfectly {smile/no joke}

I.E., I expect -- as a searcher -- that
the FRESHEST data will be be used to
calculate search results. (A two-week return
to older data once new data is "in play"
seems ... well .. not what I expect.)

BOTTOM LINE:
Is this reversion to older cached snapshots
"normal" ... "unheard of" ... or what?

[nuevojefe]

I believe it is just that the datacenters have different data sets and the results on www for instance are seen slightly different if the datacenter being used is still rotating.

IMHO

[steveb]

Normal.

Some of their master caches continue to be many months old.

[kaled]

Fresh data seems to be bodged into the index rather than properly integrated. If your pages change every couple of days and Googlebot reads them then the cache may stay uptodate. However, once a page is a few days old, the cache seems to revert back to the one in the main index.

Kaled.

[nuevojefe]

Not that it will address your specific incident, but a good way to tell is to place the date and time or something similar on your pages so that you know when the cache is from.

[Meta_Vision]

Many thanks to all.

I suspected this was "normal" ...

... BUT I am not pleased.
{smile/no joke}

TO BE CLEAR: Google is the search engine I use
almost exclusively for my own searches,
AND it is usually Google that has the quickest
update of changes to my sites. I'm a fan.

HOWEVER:
As much as I understand (conceptually) the
technical difficulty of synchronizing the
(distributed) Google database it bothers me ..

   -- both as a searcher and a webmaster --

... that it is "normal" for OLD INFORMATION
to "usurp" FRESH.

Well, I guess that pretty much wraps this up.
(But if anyone has more to say, please do.)
Again, my thanks to all for your quick response.

BOTTOM LINE: {serious smile}
I hope the magnificent Google Team
will -- oh so elegantly -- erase this
searcher's concern about FRESHNESS....

...someday... soon. # # #

[4crests]

Meta_Vision...

I'm experiencing the same thing. I have several Yahoo Stores, and they usually get visited every day. Usually the cache will show my site from 1 or 2 days prior. Around the first of this month, my page suddenly reverted to a cache of January 22nd. I haven't seen an updated cache for my page since then.

It is interesting to note, that ALL 4 of my Yahoo Stores all did the same thing, and all are stuck on a Cache date of Jan. 22nd. Also, I have several good friends who all have Yahoo Stores, and they too are stuck on a cache date of Jan. 22nd.

[Meta_Vision]

Thanks, 4crests, for that date (1/22)

Mine reverted to roughly the same
time period as yours is hung. May not
mean anything {smile} but still good to
have that data point.

    (I've added an update date to the page
    since then so I can tell precisely from
    now on. thanks, nuevojefe)

# # #

[Meta_Vision]

"BIG PICTURE" NOTE ...

I continue to come across "stuck cache"
messages in forums (one said explictly Jan 21)
and -- for many -- this "sticking" is unusual....

... BUT the reason I've come back to this thread
is to say something that has just come to
mind about this.

IF Google is going to claim 4 billion+ pages ...

.. it will have to keep it's indexes FRESHER,

Or that claim is "meaningless."

Searching through "old data" in a fast changing world is not what I expect.

To say Google keeps fresh data from "the most
popular sites" ... only tells me I can get the latest
news on Britney ...

                          ... but not the most significant
new data in a field of study that may be -- for now --
too obscure to be "popular." The information that
could change the world ... may have to wait for
indexing until what? Popularity? {guess my expression}

This is a bias of the past over the future

I am a searcher first.

These "stuck cache" stories tell me
Google is not providing what I'm expecting ...

              The freshest information

                    from sites large and tiny --

                   popular or nearly unknown.

Find a solution, sweet riders of complexity {smile} ...

... the searcher is always right. AMEN

# # #

[kaled]

It is wholly unrealistic to expect any search engine to index the entire web every week, or whatever. Priorities have to be set. Whilst Google may not be rated 10/10 by everyone in regard to this, a dispassionate obvserver would probably say, overall, they do a pretty good job.

Kaled.

[Meta_Vision]

"unrealistic" is a backward-looking word ...

Creative technological engineering ...

                                           makes the future.

The searcher is always right.
{serious smile}

[k18boy]

(first time caller!)

i dont think anyone expects google to index the entire web every week... but just to keep the most resent cache!

has anyone asked gg why the cache goes back in time?

[Stefan]

There is a default cache that shows in the serps between fluxing fresh results. It updates about every 4 weeks for a PR6 index.htm site. This is what you're seeing. It's normal.

Over the last few weeks, the percentage of the time that the default, non-fresh cache, has been showing up in the serps has been much less that it was before. Google is very nimble lately, (and it sure beats hell out of the new Yahoo...)

[steveb]

Staring at a PR6 page with a cache from September that has been crawled 30+ times since then. Maybe it is domain-centered somehow but reverting to the cromagnon caches half the time is not good. I also checked the PR5 one I know that keeps reverting to last August, and its still there (been crawled only six to ten times though).

===

It shows the old cache half the time, and the newest crawl cache the other half.

[edited by: steveb at 3:17 am (utc) on Feb. 20, 2004]

[Stefan]

Staring at a PR6 page with a cache from September that has been crawled 30+ times since then.

Does it never show a fresh cache?

I'm extrapolating from my own site... guess it's different for others. I get older page caches, fluxing in and out in between fresh results. They revert to a default non-fresh cache that seems to be updated about every 4 weeks.

Perhaps if you stare at it less it will go away... :-)

[4crests]

happy to say, mine finally updated a couple days ago, and seems to be back on a daily schedule.