Greg Stark wrote:
> It looks like the row *was* updated by transaction 6179 and the new
> version was stored in line pointer 12. However it's marked
> XMAX_INVALID which means at least somebody at some point thought 6179
> had aborted and marked that hint bit.
Hmm, but LP 12 shouldn't have len=0 then, no? Unless it has been
clobbered by vacuum or HOT pruning, I guess ...
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.
In response to
pgsql-bugs by date
|Next:||From: Sebastien Lardiere||Date: 2009-08-21 16:26:11|
|Subject: BUG #5004: pg_freespacemap make a SegFault|
|Previous:||From: Alvaro Herrera||Date: 2009-08-21 16:15:26|
|Subject: Re: 8.4.0 data loss / HOT-related bug|