<p dir="ltr">On Jun 13, 2013 4:18 AM, "Stephen Frost" <<a
href="mailto:sfrost@snowman.net">sfrost@snowman.net</a>>wrote:<br /> > * Ants Aasma (<a
href="mailto:ants@cybertec.at">ants@cybertec.at</a>)wrote:<br /> > > In a cluster setting you take the CSN value
onthe master, then before<br /> > > starting execution on a standby you wait until that the standby has<br />
>> replayed enough WAL to reach the CSN point read from the master and<br /> > > you know that after that
everythingthat the snapshot can see is also<br /> > > replayed on the standby.<br /> ><br /> > This does
makea lot of sense- but to clarify, this would only be for<br /> > certain isolation levels, right? Or would we
implementthis for every<br /> > snapshot taken in a read-committed transaction?<p dir="ltr">I don't see a way how
snapshotsrepresenting different points in time could provide sensible results for parallel queries, so this needs to be
usedfor all snapshots. This is why having the capability to request for a snapshot that is fresh enough for a specific
clientbut old enough to not require replication waits would be a good feature.<p dir="ltr">Regards,<br /> Ants Aasma