Horton

Re: WAL format changes

Peter Eisentraut <peter_e@gmx.net>

View thread
From: Peter Eisentraut <peter_e@gmx.net>
To: Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>
Cc: PostgreSQL-development <pgsql-hackers@postgreSQL.org>
Date: 2012-07-06T21:06:58Z
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 →

  1. Update pgindent install instructions and update typedef list.

On fre, 2012-06-15 at 00:01 +0300, Heikki Linnakangas wrote:
> 1. Use a 64-bit segment number, instead of the log/seg combination. And 
> don't waste the last segment on each logical 4 GB log file. The concept 
> of a "logical log file" is now completely gone. XLogRecPtr is unchanged, 
> but it should now be understood as a plain 64-bit value, just split into 
> two 32-bit integers for historical reasons. On disk, this means that 
> there will be log files ending in FF, those were skipped before.

A thought on this.  There were some concerns that this would silently
break tools that pretend to have detailed knowledge of WAL file
numbering and this previous behavior of skipping the FF files.  We could
address this by "fixing" the overall file naming from something like

00000001000008D0000000FD
00000001000008D0000000FE
00000001000008D0000000FF
00000001000008D100000000

to

00000001000008D0FD000000
00000001000008D0FE000000
00000001000008D0FF000000
00000001000008D100000000

which represents the new true WAL stream numbering as opposed to the old
two-part numbering.

Thus, any tool that thinks it knows how the WAL files are sequenced will
break very obviously, but any tool that just looks for 24 hexadecimal
digits will be fine.

I wonder if any tools in the former category would also break if one
changes XLOG_SEG_SIZE.