Christopher Browne <cbbrowne at ca.afilias.info> writes:
>> OK. Will do early next week.

> I think it warrants having some more discussion of what the shape of
> policy for this ought to be.

I had a further chat with Jan, and it points at some further ideas
about policy...

> The right answer may be for the interval (currently 10 minutes) to be
> a parameter passed by slon to cleanupEvent.

That's one piece that would seem worthwhile.

It seems as though we might want two new parameters to this.

1.  Give the user the option of continuing to use the DELETE logic as
    it stands now.  Choice seems like a good thing.

2.  The interval value that presently is a constant value function
    cleanupEvent() should become a slon parameter, to allow the
    user to pick how long to keep old data.

    If that value is, say, 48h, then a pg_dump of a subscriber will be
    a candidate for being useful for CLONE NODE until the end of that
    48h period.

    That makes the dump a fair bit more valuable.
-- 
output = ("cbbrowne" "@" "ca.afilias.info")
<http://dba2.int.libertyrms.com/>
Christopher Browne
(416) 673-4124 (land)