Обсуждение: Show the LSN in rm_redo_error_callback

Hi,

I've repeatedly stared at logs containing PANICs during WAL
replay. Unfortunately it's rather hard to find out which record actually
caused the problem as we print the record's contents but not its LSN.

I think it's a no-brainer to add that to master. But I'd even argue that
it'd be a good idea to add it to the backbranches - there've been a
significant number of bugs causing PANICs due to replay errors in the
past (and we might be hunting one more of them right now).

Greetings,

Andres Freund

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

On 2/10/15 5:15 PM, Andres Freund wrote:
> Hi,
> 
> I've repeatedly stared at logs containing PANICs during WAL
> replay. Unfortunately it's rather hard to find out which record actually
> caused the problem as we print the record's contents but not its LSN.
> 
> I think it's a no-brainer to add that to master.

Makes sense.

> But I'd even argue that
> it'd be a good idea to add it to the backbranches - there've been a
> significant number of bugs causing PANICs due to replay errors in the
> past (and we might be hunting one more of them right now).

We've had WAL since 7.1, and this is the first I've heard of this idea,
so while I sympathize with the idea, it's hard to agree that this is a
must-have.  The standard for back-patching ought to be higher than mere
convenience.