Simon Riggs <simon(at)2ndQuadrant(dot)com> writes:
> Thus there is no guarantee that this is sufficient to have archived all
> the files you would like to archive. The patch does not provide a clean
> shutdown in all cases and since you don't know what state its in, you
> are still forced to take external action to be safe, exactly as you do
> If I didn't already say, I came up with exactly the same solution 2
> years ago and then later explained it didn't work in all cases. I'm
> saying the same thing again here now.
What's your point? Surely the applied patch is a *necessary* component
of any attempt to try to ensure archiving is complete at shutdown.
I agree that it doesn't cover every risk factor, and there are some
risk factors that cannot be covered by Postgres itself. But isn't it
a step in a desirable direction?
regards, tom lane
In response to
pgsql-hackers by date
|Next:||From: Tom Lane||Date: 2009-05-28 22:16:56|
|Subject: Re: pg_migrator and an 8.3-compatible tsvector data type |
|Previous:||From: Joshua Tolley||Date: 2009-05-28 22:07:39|
|Subject: Dtrace probes documentation|