On Mon, Dec 31, 2012 at 1:38 AM, Pavan Deolasee <pavan.deola...@gmail.com>wrote:
> On Sun, Dec 30, 2012 at 12:38 AM, Stephen Frost <sfr...@snowman.net> > wrote: > > * Pavan Deolasee (pavan.deola...@gmail.com) wrote: > >> On Fri, Sep 7, 2012 at 6:06 PM, Kevin Grittner > >> > That makes sense to me. The reason I didn't make that change when I > >> > added the serializable special case to pg_dump was that it seemed > >> > like a separate question; I didn't want to complicate an already big > >> > patch with unnecessary changes to non-serializable transactions. > >> > > >> > >> If we agree, should we change that now ? > > > > This is on the next commitfest, so I figure it deserves some comment. > > For my part- I tend to agree that we should have it always use a read > > only transaction. Perhaps we should update the pg_dump documentation to > > mention this as well though? Pavan, do you want to put together an > > actual patch? > > > > I'd posted actual patch on this thread, but probably linked wrong > message-id in the commitfest page. Will check and correct. Regarding > pg_dump's documentation, I don't have strong views on that. Whether > pg_dump runs as a read-only transaction or not is entirely internal to > its implementation, but then if we make this change, it might be worth > telling users that they can trust that pg_dump will not make any > changes to their database and hence a safe operation to carry out. > I have updated the commitfest submission to link to the correct patch email. I initially thought that this patch deserves accompanying documentation because pg_dump's serializable transaction may error out because of a conflict. But the following line in the docs [1] confirms otherwise: "read-only transactions will never have serialization conflicts" So no doc patch necessary :) [1] http://www.postgresql.org/docs/9.2/static/transaction-iso.html -- Gurjeet Singh http://gurjeet.singh.im/