Re: Process local hint bit cache - Mailing list pgsql-hackers

From Merlin Moncure
Subject Re: Process local hint bit cache
Date
Msg-id AANLkTi=4m0To4uA3eaqfTbQ+ydk9vDksy9-UPVVJ0f=U@mail.gmail.com
Whole thread Raw
In response to Re: Process local hint bit cache  (Merlin Moncure <mmoncure@gmail.com>)
Responses Re: Process local hint bit cache
Re: Process local hint bit cache
List pgsql-hackers
On Wed, Mar 30, 2011 at 2:35 PM, Merlin Moncure <mmoncure@gmail.com> wrote:
>
> btw I haven't forgotten your idea to move TransactionIdInProgress
> Down. I think this is a good idea, and will experiment with it pre and
> post cache.

aside:
Moving TransactionIdInProgress below TransactionIdDidCommit can help
in once sense: TransactionIdDidCommit grabs the XidStatus but discards
the knowledge if the transaction is known aborted.   If it is in fact
aborted you can immediately set the hint bits and punt. This should
save an awful lot of calls to TransactionIdInProgress when scanning
unhinted dead tuples.

otoh, checking commit status first can burn you if you are repeatedly
tuples that are in transaction, especially your own.  Probably this
could be mitigated by splitting TransactionIdIsInProgress so that you
can do just the local checks and not shared memory, so that you could:
1. is this transaction mine? (if so, we are done)
2. get xid status if commit/abort done
3. do ProcArray portions of TransactionIdIsInProgress

merlin


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: Problem with pg_upgrade?
Next
From: Robert Haas
Date:
Subject: Re: Problem with pg_upgrade?