>>Uh... no. The whole point of doing things in shared buffers is that
>>you don't have to write and fsync the buffers immediately. Instead,
>>buffer evicting handles that stuff for you.
So you mean to say that there exists operations where Xlog is not required
even though it marks the buffer as dirty for later eviction.
>>I don't see why it's any different from log_newpage() in that regard.
>>That data is initialized before being written, as well, but someone
>>contemplated the possible need to write a page of all zeros.
The comment above the code indicates that "the page is uninitialized".
Does the code consider page with all zero's as uninitialized or the comment
is not
appropriate.
-----Original Message-----
From: Robert Haas [mailto:robertmhaas@gmail.com]
Sent: Sunday, June 10, 2012 7:14 PM
To: Amit kapila
Cc: Tom Lane; pgsql-hackers@postgresql.org
Subject: Re: [HACKERS] log_newpage header comment
On Sat, Jun 9, 2012 at 1:43 AM, Amit kapila <amit.kapila@huawei.com> wrote:
>>On further review, I think that we ought to make MarkBufferDirty() the
>>caller's job, because sometimes we may need to xlog only if
>>XLogIsNeeded(), but the buffer's got to get marked dirty either way.
>
> Incase the place where Xlog is not required, woudn't it fsync the data;
> So in that case even MarkBufferDirty() will also be not required.
Uh... no. The whole point of doing things in shared buffers is that
you don't have to write and fsync the buffers immediately. Instead,
buffer evicting handles that stuff for you.
>>So I think the new function should just do step 5 - emit XLOG and set
>>LSN/TLI.
>
> In the API log_newpage_buffer(), as buffer already contains the page to be
logged, so can't it be assumed that the page will be initialized and no need
to check
> if PageIsNew before setting LSN/TLI.
I don't see why it's any different from log_newpage() in that regard.
That data is initialized before being written, as well, but someone
contemplated the possible need to write a page of all zeros.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
