On Thu, Sep 13, 2012 at 1:45 PM, Jeff Davis pg...@j-davis.com wrote:
On Thu, 2012-09-13 at 12:39 -0400, Robert Haas wrote:
On Wed, Sep 12, 2012 at 7:19 PM, Jeff Davis pg...@j-davis.com wrote:
This bug seems particularly troublesome because the right fix would be
to include the relpersistence
On Wed, Sep 12, 2012 at 7:19 PM, Jeff Davis pg...@j-davis.com wrote:
This bug seems particularly troublesome because the right fix would be
to include the relpersistence in the WAL records that need it. But that
can't be backported (right?).
No, because if a WAL record was written at all, then
On Thu, 2012-09-13 at 12:39 -0400, Robert Haas wrote:
On Wed, Sep 12, 2012 at 7:19 PM, Jeff Davis pg...@j-davis.com wrote:
This bug seems particularly troublesome because the right fix would be
to include the relpersistence in the WAL records that need it. But that
can't be backported
Indeed, this is a nasty bug that leads to data corruption. The following
sequence results in corruption of the visibility map, but I believe it
can be shown to cause problems for a btree or GIN index as well. So it's
recoverable if you do a VACUUM or a reindex.
drop table foo;
create table foo(i