Re: Transaction Snapshot Cloning

simon@2ndquadrant.com (Simon Riggs) writes:
> On Fri, 2008-01-11 at 20:39 +0000, Simon Riggs wrote:
>> On Fri, 2008-01-11 at 15:05 -0500, Tom Lane wrote:
>> Simon Riggs <simon@2ndquadrant.com> writes:
>> > > If we had a function 
>> > >     replace_serializable_snapshot(master_xid, txid_snapshot)
>> > > this would allow us to use the txid_snapshot values to replace our
>> > > transaction's serializable snapshot.
>> > 
>> > ... whereupon we'd get wrong answers.  Certainly you could not allow
>> > transaction xmin to go backwards, and I'm not sure what other
>> > restrictions there would be, but the whole thing gives me the willies.
>
> Sorry, forgot to add
> - global xmin isn't going backwards
> - neither is latest completed xid
>
> The xmin of the transaction will go backwards, but as long as we don't
> do anything prior to the setting of the cloned snapshot, what can go
> wrong? :-)

Note that we required that the "provider transaction" have the
attributes IsXactIsoLevelSerializable and XactReadOnly both being
true, so we have the mandates that the resultant backend process:

a) Is in read only mode, and
b) Is in serializable mode.

That's a pair of (possibly stretching-wide!) suspenders worth of
support from Evil...
-- 
(format nil "~S@~S" "cbbrowne" "cbbrowne.com")
http://linuxfinances.info/info/internet.html
Trying  to be happy  is like trying to build   a machine for which the
only specification is that it should run noiselessly.