Re: Odd transaction timestamp sequence issue - Mailing list pgsql-general

From Tom Lane
Subject Re: Odd transaction timestamp sequence issue
Date
Msg-id 10296.1145633799@sss.pgh.pa.us
Whole thread Raw
In response to Re: Odd transaction timestamp sequence issue  (Jeff Amiel <jamiel@istreamimaging.com>)
List pgsql-general
Jeff Amiel <jamiel@istreamimaging.com> writes:
> I thought:
> "Each transaction sees a snapshot (database version) as of its
> starttime, no matter what other transactions are doing while it runs"

That's a correct statement in SERIALIZABLE mode, but in the default
READ COMMITTED mode, it's more complicated --- a new snapshot is taken
for each command within a transaction.  See the docs.

            regards, tom lane

pgsql-general by date:

Previous
From: Alban Hertroys
Date:
Subject: Re: A few questions about ltree
Next
From: Stephan Szabo
Date:
Subject: Re: A few questions about ltree