0x44

Поиск
Список
Период
Сортировка
От Moray McConnachie
Тема 0x44
Дата
Msg-id 00ca01bf4b0e$25490130$760e01a3@oucs.ox.ac.uk
обсуждение исходный текст
Список pgsql-general
I've very recently experienced a failure of my postgres database which
neither vacuuming nor rebuilding indices cured.  This may have been
caused by my being forced to flip the switch - something that's never
happened to me under linux before, and on a machine with a light load
running nothing but postmaster, some postgres connections, a small
selection of admin demons and sshd.

Thereafter vacuum would die processing certain tables, and pg_dump
would fail similarly. Certain simple operations under psql failed
also, e.g. 'select fieldone,fieldtwo from table;'

I'm afraid I simply restored from backups, so I'm not in a position to
speculate on why it might have occurred, since I didn't delve much
further.

However, the message with which postgres crashed was along the lines
of

"backend received a 0x44 while idle" or something like that. This
appears to come from fe-exec.c, i.e. libpq, presumably.

Can anyone tell me what this means in the context? I did look at
source, but my C is non-existent.

I'm keeping a close eye out now, and will obviously post with logs if
the event recurs in the future.
Thanks,
Moray

----------------------------------------------------------------------
----------------
Moray.McConnachie@computing-services.oxford.ac.uk


В списке pgsql-general по дате отправления:

Предыдущее
От: "Gurunandan R. Bhat"
Дата:
Сообщение: Trying to understand SPI
Следующее
От: Thomas Reinke
Дата:
Сообщение: Re: [GENERAL] copy command -- foiled by pg_atoi