Re: pitr question

Поиск
Список
Период
Сортировка
От Craig Ringer
Тема Re: pitr question
Дата
Msg-id 4CB67A5D.6090300@postnewspapers.com.au
обсуждение исходный текст
Ответ на Re: pitr question  ("Joshua D. Drake" <jd@commandprompt.com>)
Список pgsql-general
On 14/10/10 00:18, Joshua D. Drake wrote:
> On Wed, 2010-10-13 at 11:40 -0400, Geoffrey Myers wrote:
>> On 10/13/2010 11:30 AM, zhong ming wu wrote:
>>> On Wed, Oct 13, 2010 at 11:17 AM, Geoffrey Myers
>>> <lists@serioustechnology.com <mailto:lists@serioustechnology.com>> wrote:
>>>  > Excuse the ignorance, but I see the following in the docs:
>>>  >
>>>  > 'In any case the hardware architecture must be the same — shipping from,
>>>  > say, a 32-bit to a 64-bit system will not work.'
>>>
>>> postgres only sees the OS. It does not care what hardware you got. So
>>> you must use the same type of OS: both 32 bit or both 64 bit.
>>
>> As I expected, thus the document should possibly be rephrased to
>> indicate it's the OS not the physical hardware that is the issue.
>
> No. Because PostgreSQL itself must also be the same architecture.

No it doesn't. As Tom Lane pointed out, plenty plenty of 64-bit OSes on
64-bit hardware will run 32-bit binaries. Additionally, most 64-bit
hardware will run 32-bit OSes with 32-bit binaries.

So long as Pg has the same word sizes, endianness, etc, it won't care.
This means that you can't log-ship from 32-bit Intel to 32-bit
big-endian PPC and expect it to work, but you can log-ship from 32-bit
Pg running on 64-bit Intel to 32-bit Pg running on 32-bit Intel and
that'll be fine.

--
Craig Ringer

Tech-related writing: http://soapyfrogs.blogspot.com/

В списке pgsql-general по дате отправления:

Предыдущее
От: Rajesh Kumar Mallah
Дата:
Сообщение: Re: Gripe: bytea_output default => data corruption
Следующее
От: Pavel Stehule
Дата:
Сообщение: Re: How to iterate fields into a NEW.* RECORD in a TRIGGER procedure ?