Re: WALInsertLock tuning
Simon Riggs <simon@2ndquadrant.com>
From: Simon Riggs <simon@2ndQuadrant.com>
To: Jeff Janes <jeff.janes@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2011-06-06T22:05:59Z
Lists: pgsql-hackers
Commits
Same data as JSON:
GET /api/v1/messages/:b64id/commits
the thread's linked commits as JSON, with link sources.
API reference →
-
Introduce compact WAL record for the common case of commit (non-DDL).
- 465883b0a2b4 9.2.0 cited
On Mon, Jun 6, 2011 at 10:09 PM, Jeff Janes <jeff.janes@gmail.com> wrote: > On Mon, Jun 6, 2011 at 11:27 AM, Simon Riggs <simon@2ndquadrant.com> wrote: >> >> But that even assumes we write the unzeroed data at the end of the >> buffer. We don't. We only write data up to the end of the WAL record >> on the current page, unless we do a continuation record, > > I see no codepath in XLogWrite which writes anything other than full > buffer pages. Second line, boolean variable called "ispartialpage". As I mentioned, even if spare bytes at the end of page were written, they aren't ever read except in corner case bugs that would be trapped by other logic put there to protect us. We're safe. If I didn't believe it and hadn't tested it, I wouldn't speak. -- Simon Riggs http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services