On Thu, Jun 28, 2018 at 1:39 AM, Michael Paquier <michael@paquier.xyz> wrote:
> On Thu, Jan 11, 2018 at 10:35:22PM -0500, Stephen Frost wrote:
>> Magnus, this was your thread to begin with, though I know others have
>> been involved, any chance you'll be able to review this for commit
>> during this CF? I agree that this is certainly a good thing to have
>> too, though I've not looked at the patch itself in depth. Is there
>> anything we can do to help move it along?
>
> As an effort to move on with bug items in the commit fest, attached are
> two patches with a proposed commit message as well as polished comments
> Those are proposed for a back-patched. The 2PC issue is particularly
> bad in my opinion because having any 2PC file on-disk and corrupted
> means that a transaction is lost. I have been playing a bit with
> hexedit and changed a couple of bytes in one of them... If trying to
> use a base backup which includes one, then the standby reading it would
> be similarly confused.
Thanks to Michael for progressing this.
Back in August, nearly a year ago, Robert Haas said upthread:
> This bug fix has been pending in "Ready for Committer" state for about
> 4.5 months. Three committers (Magnus, Heikki, Tom) have contributed
> to the thread to date. Maybe one of them would like to commit this?
Although the state is now back to "Needs Review", I echo those
sentiments. This issue has now been hanging around for about 18
months.
cheers
andrew
--
Andrew Dunstan https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services