> Also, for Mario and Barry: does this test case look anything like what
> your real applications do? In particular, do you ever do a SELECT FOR
> UPDATE in a transaction that commits some changes, but does not update
> or delete the locked-for-update row? If not, it's possible there are
> yet more bugs lurking in this area.
>
> regards, tom lane
I've checked the application, when I select for update I will update those tuples, though it might be an
update where no real modification is done (e.g. update table set col1=col1).
I'm pretty sure I've identified the source of the problem in my application, but in this specific place there
is no "select for update", but a rollback while another update is in progress. I guess this is triggering
the problem now and then.
But for the scenario you mention above, I cannot imagine how this might happen in my application, it's not
easy to say for sure, it's a quite complex web based content management system and not easy to debug such
errors, because I've no clue how to trigger it reproduceable.
Best regards,Mario Weilguni
