Re: Dividing progress/debug information in pg_standby, and stat before copy

On 1/26/10 10:58 AM, Tom Lane wrote:
> Josh Berkus <josh(at)agliodbs(dot)com> writes:
>> On 1/26/10 10:48 AM, Heikki Linnakangas wrote:
>>> I didn't intend to replace pg_standby when I started this, it just kind
>>> of happened. Maybe we should provide a sample script similar to
>>> pg_standby, to be used instead of plain 'cp', that does the cleanup too.
>
>> What I'm pointing out is that you haven't *quite* replaced pg_standby,
>> and at this late date aren't going to. Some people will still want to
>> use it. Especially for people who for some reason aren't using SR.
>
> Right, but the question is: is there enough use-case left in it to
> justify spending community effort on polishing rough edges? It's not
> like we haven't got plenty else to do to get 9.0 out the door.

Can we wait until the alpha to decide that? I haven't tested Heikki's
code to find out whether it works as he intends, and Simon seems to
think different. If it's OK to wait until beta to fix the error
messages for pg_standby, then let's wait.

Anyway, if we fix the *core* bogus error messages for standby mode, that
will eliminate half of what's confusing and alarming people (and all of
it for those using SR).

--Josh Berkus