Re: amcheck assert failure

On Mon, Apr 22, 2019 at 9:03 AM Grigory Smolkin
<g.smolkin@postgrespro.ru> wrote:
> Wow, that` neat, thank you for a dump instruction!
> Attached.

This looks like it was deliberately corrupted. Line pointer 255 is an
LP_REDIRECT, line pointer, while line pointer 256 and 257 are
LP_UNUSED, which should never happen.

The ascii string "blablahblah" has been overwritten at byte offset
1040 (from the beginning of the page). I would take a closer look at
what those tests do. Either the problem is in the tests, or your
database is haunted, which is outside my area of expertise.

-- 
Peter Geoghegan