Re: svn commit: r1483795 - /subversion/branches/1.8.x/STATUS

On Sat, May 18, 2013 at 12:55 AM, Bert Huijben <bert_at_qqmail.nl> wrote:
>> -----Original Message-----
>> From: Ivan Zhakov [mailto:ivan_at_visualsvn.com]
>> Sent: vrijdag 17 mei 2013 17:03
>> To: C. Michael Pilato
>> Cc: Philip Martin; Branko Čibej; dev_at_subversion.apache.org
>> Subject: Re: svn commit: r1483795 - /subversion/branches/1.8.x/STATUS
>>
>> On Fri, May 17, 2013 at 6:59 PM, C. Michael Pilato <cmpilato_at_collab.net>
>> wrote:
>> > On 05/17/2013 10:55 AM, Philip Martin wrote:
>> >> Branko Čibej <brane_at_wandisco.com> writes:
>> >>
>> >>> On 17.05.2013 15:32, ivan_at_apache.org wrote:
>> >>>> --- subversion/branches/1.8.x/STATUS (original)
>> >>>> +++ subversion/branches/1.8.x/STATUS Fri May 17 13:32:56 2013
>> >>>> @@ -124,6 +124,14 @@ Candidate changes:
>> >>>> Votes:
>> >>>> +1: stefan2 (for 1.8.1)
>> >>>>
>> >>>> +* r1483781
>> >>>> + Fix FSFS repository corruption on power or network disk failure on
>> Windows:
>> >>>> + http://svn.haxx.se/dev/archive-2013-05/0245.shtml
>> >>>> + Justification:
>> >>>> + Repository corruption. Regression from 1.6.x
>> >>>> + Votes:
>> >>>> + +1: ivan
>> >>>> +
>> >>>
>> >>> Is this considered a blocker? Should we roll RC3 next week and restart
>> >>> the soak period?
>> >>
>> >> We should put this into 1.8.0 but I don't think it is a destabilizing
>> >> change so we don't need to restart the soak.
>> >
>> > I agree. Soak time extensions are tied to the complexity of the change, not
>> > the severity of the bug fixed.
>> >
>> I agree this change is pretty simple and actually it just reverts
>> Subversion to 1.6.x behavior, but there are other places with similar
>> issue and they may require more complicated fix. I'm working on them.
>
> If this and the future followups are going to have a huge performance impact we
> should probably make the full fsync option configurable for those who have a battery
> backed up storage.
Yes, this will be good improvement anyway, but I think repository
integrity should be first goal.

Also please note that FlushFileBuffers doesn't perform *full* sync: it
flushes data only for one file.

-- 
Ivan Zhakov
CTO | VisualSVN | http://www.visualsvn.com