Re: bytea operator bugs (was Re: [GENERAL] BYTEA, indexes
От | Joe Conway |
---|---|
Тема | Re: bytea operator bugs (was Re: [GENERAL] BYTEA, indexes |
Дата | |
Msg-id | 3D610AD3.9030507@joeconway.com обсуждение исходный текст |
Ответ на | bytea operator bugs (was Re: [GENERAL] BYTEA, indexes and "like") (Joe Conway <mail@joeconway.com>) |
Ответы |
Re: bytea operator bugs (was Re: [GENERAL] BYTEA, indexes
|
Список | pgsql-patches |
Tom Lane wrote: > Joe Conway <mail@joeconway.com> writes: > >>parts=# SELECT * from bombytea where parent_part like '02-%'; >>server closed the connection unexpectedly > > You should stop right here and investigate the reason for the crash. It's the same reason as before -- i.e. the TEXTOID assertion in patternsel(). I just didn't expect to see physical index corruption as a result. >>So it appears that "SELECT * from bombytea where parent_part like >>'02-%'; server closed the connection unexpectedly" somehow physically >>corrupts the index! > > Hm, I'd have expected WAL recovery to fix that. But the easiest avenue > of attack is to isolate the reason for the crash. Please post a test > case if you want help. OK. I'll be back at this again this morning on a different development machine. I'll see if I can get a reliable test case and post it. Joe
В списке pgsql-patches по дате отправления: