On 02/12/14 20:21, Andres Freund wrote:
> On 2014-11-24 13:16:24 +0200, Heikki Linnakangas wrote:
>> To be clear: I don't think this API is very good for its stated purpose, for
>> implementing global sequences for use in a cluster. For the reasons I've
>> mentioned before. I'd like to see two changes to this proposal:
>> ...
>> 2. Instead of the single amdata field, make it possible for the
>> implementation to define any number of fields with any datatype in the
>> tuple. That would make debugging, monitoring etc. easier.
>
> My main problem with that approach is that it pretty much nails the door
> shut for moving sequences into a catalog table instead of the current,
> pretty insane, approach of a physical file per sequence. Currently, with
> our without seqam, it'd not be all that hard to force it into a catalog,
> taking care to to force each tuple onto a separate page...
>
I don't know, I think if we decide to change storage format we can do
serialization/conversion in seqam layer, it does not seem to matter too
much if the serialization into some opaque type is done in AM itself or
by the API layer. Or we can have one relation for all sequences in
single AM, etc. In general I don't think that the custom columns for AM
approach prohibits future storage changes.
-- Petr Jelinek http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training &
Services