On Fri, Mar 20, 2026 at 12:26:36PM +0000, Ana Almeida wrote:
> We executed DELETE, VACUUM, and REINDEX commands multiple
> times. During one of the executions, the REINDEX operation failed
> with the error: "could not open file". After this, we dropped and
> recreated the database and repeated the tests. In one of the
> executions, the same REINDEX command resulted in segmentation fault,
> which caused the database to crash. When the REINDEX command failed,
> it left behind temporary index copies with the _ccnew suffix. We
> manually dropped these indexes and re-ran the REINDEX command, which
> then completed successfully.
I am afraid that it is one of these cases where having a
self-contained workload to be able to reproduce the issue, even if the
issue can be hit at a very low rate, would be super useful before we
could categorize that as a backend core issue. I could buy that there
is an in-core problem, but it's hard to justify the time investment
based on an assumption that we may have a problem.
FWIW, I have not seen such problematic error patterns lately, so it's
hard to say. I'll double-check around me and see if there is some data
to fish. Perhaps there is something matching with a portion of your
problematic patterns, or that has a similar smell.
--
Michael
