WALInsertLock tuning

Simon Riggs <simon@2ndquadrant.com>

From: Simon Riggs <simon@2ndQuadrant.com>

To: pgsql-hackers <pgsql-hackers@postgresql.org>

Date: 2011-06-06T16:15:28Z

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

Attachments

walinsertlock_avoid_zero.v1.patch (application/octet-stream) patch v1

In earlier discussions of how to improve WALInsertLock contention, it
was observed that we must zero each new page before we advance the WAL
insertion point.
http://postgresql.1045698.n5.nabble.com/Reworking-WAL-locking-td1983647.html

IMHO the page zeroing is completely unnecessary, and replication works
perfectly well without that (as a test of recovery logic). It is
unnecessary because we already allow non-zeroed parts of WAL files,
and provide a mechanism to detect stale data.

The following trivial patch removes the page zeroing, which reduces
the lock duration.

Comments?

-- 
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services