Re: SSI atomic commit

From: "Kevin Grittner" <Kevin(dot)Grittner(at)wicourts(dot)gov>
To: "Heikki Linnakangas" <heikki(dot)linnakangas(at)enterprisedb(dot)com>
Cc: <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: SSI atomic commit
Date: 2011-07-05 18:15:13
Message-ID: 4E130E61020000250003EFD8@gw.wicourts.gov
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> wrote:

> Hmm, I think it would be simpler to decide that instead of
> SerializableXactHashLock, you must hold ProcArrayLock to access
> LastSxactCommitSeqNo, and move the assignment of commitSeqNo to
> ProcArrayTransaction(). It's probably easiest to move
> LastSxactCommitSeqno to ShmemVariableCache too. There's a few
> places that would then need to acquire ProcArrayLock to read
> LastSxactCommitSeqno, but I feel it might still be much simpler
> that way.

We considered that. I think the biggest problem was that when there
is no XID it wouldn't be covered by the lock on assignment. We
couldn't see a good way to increment and assign the value without LW
lock coverage, and we didn't want to add LW locking to that code
path. If you can see a way around that issue, I agree it would be
simpler.

-Kevin

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Brar Piening 2011-07-05 18:15:25 Re: Review of VS 2010 support patches
Previous Message Heikki Linnakangas 2011-07-05 18:14:30 Re: SSI 2PC coverage