bug with vacuum analyze?

Поиск
Список
Период
Сортировка
От Trever L. Adams
Тема bug with vacuum analyze?
Дата
Msg-id 1047320988.1057.0.camel@aurora.localdomain
обсуждение исходный текст
Ответы Re: bug with vacuum analyze?
Список pgsql-bugs
I believe I have found a bug in the vacuum procedures.  I am helping out
a friend who is running a "alumni" sites of sorts using postgresql 7.2.3
as the backend.  The first page someone loads does most of the queries
that any page will need and stores it in a session file (php style) and
hands a cookie to the user.

If I never do a vacuum analyze, it takes between 8-11 (high spikes
around 23) seconds most of the time for this first page to load.  If I
do vacuum, nothing seems to change.  If I do a vacuum analyze, the
minimum load time is between 41-44 seconds.  Higher system loads,
instead of adding 4-10 seconds, easily double the number up beyond 80
seconds.  Yes, each of these tables have at least one index (the most
indexes I believe are 4).  Those with multiple, I believe, all have a
primary index key.

Effects of vacuum full and vacuum freeze haven't been tested.  The only
way to recover from these horrible times seems dump/drop/reload.

Is this a bug, known or otherwise, are their workarounds besides don't
do it?

Thank you,
Trever Adams

P.S. This is a RedHat 8.0 box with all errata fixes.  Specifics can be
provided later if needed.
--
"Perilous to all of us are the devices of an art deeper than we possess
ourselves." -- Gandalf the White [J.R.R. Tolkien, "The Two Towers", Bk
3, Ch. XI]

В списке pgsql-bugs по дате отправления:

Предыдущее
От: Bruce Momjian
Дата:
Сообщение: Re: Bug #907: psql 7.3.2 \d , \lmeta commands on windows
Следующее
От: Dan Langille
Дата:
Сообщение: man pg-dump + triggers