Re: Proposed WAL changes

On Wed, Mar 07, 2001 at 12:03:41PM -0800, Mikheev, Vadim wrote:
> Ian wrote:
> > > I feel that the fact that
> > > 
> > > WAL can't help in the event of disk errors
> > > 
> > > is often overlooked.
> > 
> > This is true in general.  But, nevertheless, WAL can be written to
> > protect against predictable disk errors, when possible.  Failing to
> > write a couple of disk blocks when the system crashes 

or, more likely, when power drops; a system crash shouldn't keep the
disk from draining its buffers ...

> > is a reasonably predictable disk error.  WAL should ideally be 
> > written to work correctly in that situation.
> 
> But what can be done if fsync returns before pages flushed?

Just what Tom has done: preserve a little more history.  If it's not
too expensive, then it doesn't hurt you when running on sound hardware,
but it offers a good chance of preventing embarrassments for (the 
overwhelming fraction of) users on garbage hardware.

Nathan Myers
ncm@zembu.com