Re: Failure while inserting parent tuple to B-tree is not fun

On 2013-10-22 16:38:05 -0400, Tom Lane wrote:
> Andres Freund <andres(at)2ndquadrant(dot)com> writes:
> > On 2013-10-22 15:24:40 -0400, Tom Lane wrote:
> >> No, that's hardly a good idea. As Heikki says, that would amount to
> >> converting an entirely foreseeable situation into a PANIC.
>
> > But IIUC this can currently lead to an index giving wrong answers, not
> > "just" fail at further inserts?
>
> No. It's exactly the same situation as when the insert is still in
> progress, ie searches have to move right when following the original
> downlink. Go read the nbtree README.

If the parent insert is still in progress we have a write lock on the
left and right page preventing such issues, right? At least that's what
the nbtree README says...
That doesn't hold true if we split a page but fail before re-inserting
the parent since the transaction rollback will obviously have released
the locks.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services