Re: Unarchived WALs deleted after crash

From: Daniel Farina <daniel(at)heroku(dot)com>
To: Heikki Linnakangas <hlinnakangas(at)vmware(dot)com>
Cc: Simon Riggs <simon(at)2ndquadrant(dot)com>, Fujii Masao <masao(dot)fujii(at)gmail(dot)com>, Jehan-Guillaume de Rorthais <jgdr(at)dalibo(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Unarchived WALs deleted after crash
Date: 2013-02-21 08:53:36
Message-ID: CAAZKuFbtsDvtzpkyAV24Us_D9w88VbvzFj9MW_1-KA-A0+gAkQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thu, Feb 21, 2013 at 12:39 AM, Heikki Linnakangas
<hlinnakangas(at)vmware(dot)com> wrote:
> On 21.02.2013 02:59, Daniel Farina wrote:
>>
>> On Fri, Feb 15, 2013 at 9:29 AM, Simon Riggs<simon(at)2ndquadrant(dot)com>
>> wrote:
>>>
>>> On 15 February 2013 17:07, Heikki Linnakangas<hlinnakangas(at)vmware(dot)com>
>>> wrote:
>>>
>>>>> Unfortunately in HEAD, xxx.done file is not created when restoring
>>>>> archived
>>>>> file because of absence of the patch. We need to implement that first.
>>>>
>>>>
>>>>
>>>> Ah yeah, that thing again..
>>>> (http://www.postgresql.org/message-id/50DF5BA7.6070200@vmware.com) I'm
>>>> going
>>>> to forward-port that patch now, before it's forgotten again. It's not
>>>> clear
>>>> to me what the holdup was on this, but whatever the bigger patch we've
>>>> been
>>>> waiting for is, it can just as well be done on top of the forward-port.
>>>
>>>
>>> Agreed. I wouldn't wait for a better version now.
>>
>>
>> Related to this, how is this going to affect point releases, and are
>> there any lingering doubts about the mechanism of the fix?
>
>
> Are you talking about the patch to avoid restored WAL segments from being
> re-archived (commit 6f4b8a4f4f7a2d683ff79ab59d3693714b965e3d), or the bug
> that that unarchived WALs were deleted after crash (commit
> b5ec56f664fa20d80fe752de494ec96362eff520)? The former was included in 9.2.0
> already, and the latter will be included in the next point release.

Unarchived WALs being deleted after a crash is the one that worries
me. I actually presume re-archivals will happen anyway because I may
lose connection to archive storage after the WAL has already been
committed, hence b5ec56f664fa20d80fe752de494ec96362eff520.

>> This is
>> quite serious given my reliance on archiving, so unless the thinking
>> for point releases is 'real soon' I must backpatch and release it on
>> my own accord until then.
>
>
> I don't know what the release schedule is. I take that to be a request to
> put out a new minor release ASAP.

Perhaps, but it's more of a concrete evaluation of how important
archiving is to me and my affiliated operation. An acceptable answer
might be "yeah, backpatch if you feel it's that much of a rush."
Clearly, my opinion is that a gap in the archives is pretty
cringe-inducing. I hit it from an out of disk case, and you'd be
surprised (or perhaps not?) how many people like to kill -9 processes
on a whim.

I already maintain other backpatches (not related to fixes), and this
one is only temporary, so it's not too much trouble for me.

--
fdr

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2013-02-21 08:54:06 Re: Materialized views WIP patch
Previous Message Heikki Linnakangas 2013-02-21 08:39:03 Re: Unarchived WALs deleted after crash