Re: data block cache algorithm and cache hit or not

From: Chuck Hamilton <chuck_hamilton_at_yahoo.com>
Date: Fri, 28 Sep 2001 10:09:59 -0400
Message-ID: <9p20bl$g17db$1@ID-85580.news.dfncis.de>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

"Howard J. Rogers" <howardjr_at_www.com> wrote in message news:3bb47dfc_at_news.iprimus.com.au...
>
> "FaheemRao" <faheemrao_at_yahoo.com> wrote in message
> news:43b58913.0109280003.1d688d02_at_posting.google.com...
> > One important thing you are missing that when you say Commit ,
> > all the Dirt duffer(changed)in the Database buffer cache are
> > written to disk as well.
> >
> > just want to add some thing to Lawrence's answer
> >
>
> It would, however, be really nice if you added something that was
> correct. What you have posted here is about as inaccurate as it is
> possible to get. Specifically, a commit most definitely does NOT
> induce DBWR to flush anything at all out of the Buffer Cache.
>
> Apart from being 100% wrong, I guess you were close.
>
> HJR
He must be thinking of a checkpoint, not a commit. Commit flushes the log buffer and nothing else.

-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 7.0.3 for non-commercial use <http://www.pgp.com>

iQA/AwUBO7SEr2m5A4kkb4ZsEQK7+wCfXEZfXKYEc5G/H9wja6WhRB4d1mgAoPeQ Wth22qCTwqbft21Jy7MVO0Gf
=e9vZ
-----END PGP SIGNATURE----- Received on Fri Sep 28 2001 - 09:09:59 CDT