Re: Memory Leak executing small queries without closing the connection - FreeBSD

Поиск
Список
Период
Сортировка
Искать
От
Tom Lane
Тема
Re: Memory Leak executing small queries without closing the connection - FreeBSD
Дата
Msg-id
25390.1450027055@sss.pgh.pa.us
Ответ на
Список
Дерево обсуждения
Memory Leak executing small queries without closing the connection Gerhard Wiesinger <lists@wiesinger.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Gerhard Wiesinger <lists@wiesinger.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Bill Moran <wmoran@potentialtech.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Gerhard Wiesinger <lists@wiesinger.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Bill Moran <wmoran@potentialtech.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Tom Lane <tgl@sss.pgh.pa.us>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Gerhard Wiesinger <lists@wiesinger.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Bill Moran <wmoran@potentialtech.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Gerhard Wiesinger <lists@wiesinger.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Bill Moran <wmoran@potentialtech.com>
Re: Memory Leak executing small queries without closing the connection - FreeBSD Adrian Klaver <adrian.klaver@aklaver.com>
Gerhard Wiesinger  writes:
>> Mem: 7814M Active, 20G Inact, 2982M Wired, 232M Cache, 1661M Buf, 30M Free
>> Swap: 512M Total, 506M Used, 6620K Free, 98% Inuse

> OK, but why do we then get: kernel: swap_pager_getswapspace(4): failed?

Just judging from the name of the function, I would bet this is a direct
result of having only 512M of swap configured.  As Bill already pointed
out, that's a pretty useless choice on a system with 32G of RAM.  As soon
as the kernel tries to push out any significant amount of idle processes,
it's gonna be out of swap space.  The numbers you show above prove that
it is almost out of free swap already.

Also, while that 20G of "inactive" pages may be candidates for reuse,
they probably can't actually be reused without swapping them out ...
and there's noplace for that data to go.

			regards, tom lane

В списке pgsql-general по дате отправления
От: Bill Moran
Дата:
От: Gerhard Wiesinger
Дата:
FAQ