[Mageia-dev] Freeze Push: e2fsprogs

Colin Guthrie mageia at colin.guthr.ie
Thu May 3 13:10:21 CEST 2012


'Twas brillig, and Thierry Vignaud at 03/05/12 11:56 did gyre and gimble:
> On 3 May 2012 12:50, Colin Guthrie <mageia at colin.guthr.ie> wrote:
>>> Sorry, but I feel this is not needed for MGA2. In fact, systemd has
>>> changed the shutdown behaviour. The good patch would be to make him sync
>>> the hwclock at shutdown, as we have no bugs because of this sysvinit
>>> behaviour.
>>
>> There are real problems related to this as described upstream. Without a
>> proper ntp implementation doing the time sync we have *no* idea that
>> what we are doing is correct. If this is a dual boot system we could be
>> clobbering the RTC without any knowledge that what we are doing is
>> vaguely valid. So no, I will not entertain adding the sync hwclock stuff
>> back. That is broken. What is also broken is that e2fsk uses the time as
>> anything to complain about. None of the other fsck implementations do
>> this AFAIK.
>>
>> So I stand by my original assertion that the change is valid and should
>> be included.
> 
> I stand to my position: it doesn't fix any issue on Mageia.
> You don't have any bug report to prove the contraty.

I've quoted a bug report and I have several VMs in a state where the
problem exists and it's sovled by this change.

Also a problem on Fedora (this is where the fix came from in the first
place): https://bugzilla.redhat.com/show_bug.cgi?id=748920

It is a REAL problem. I cannot understand why there is such opposition
to this change?


> You I doubt you can find one since we make the Drakx
> installer prevent fsck to use time or mount count to
> cause a fsck (at least on journalized systems) for years

This is nothing to do with the installer or the filesystem
setup/configuration. In your quote from earlier:

sub disable_forced_fsck {
    my ($dev) = @_;
    run_program::run("tune2fs", "-c0", "-i0", devices::make($dev));
}

the -c option is related to mount counts and -i option is to do with
interval between checks. This is NOT related to this issue. This issue
is about when the RTC is different to the time recorded in the latest
superblock. It's nothing to do with an interval between checks.

Col

-- 

Colin Guthrie
colin(at)mageia.org
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited http://www.tribalogic.net/
Open Source:
  Mageia Contributor http://www.mageia.org/
  PulseAudio Hacker http://www.pulseaudio.org/
  Trac Hacker http://trac.edgewall.org/


More information about the Mageia-dev mailing list