[sheepdog] [PATCH 2/3] tests: add -j option for testing with journaling

MORITA Kazutaka morita.kazutaka at lab.ntt.co.jp
Thu Apr 11 07:40:40 CEST 2013


At Thu, 11 Apr 2013 12:58:43 +0800,
Liu Yuan wrote:
> 
> On 04/11/2013 12:48 PM, MORITA Kazutaka wrote:
> > Hitoshi Mitake wrote:
> >> > 
> >> > At Wed, 10 Apr 2013 16:05:54 +0800,
> >> > Liu Yuan wrote:
> >>> > > 
> >>> > > On 04/10/2013 03:51 PM, Hitoshi Mitake wrote:
> >>>> > > > I don't disagree with making journaling as a default feature. But it
> >>>> > > > seems that current journaling needs to be improved and debugged. I
> >>>> > > > believe removing -j option should be done after enough stabilizing.
> >>> > > 
> >>> > > This process looks sens to me. I agree journaling need much debugging
> >>> > > and improvement. So what do you think of 'improvement' to bring to
> >>> > > journaling before we can apply it as default?
> >> > 
> >> > I believe -j option can be removed after passing all tests with -j
> >> > option itself. Currently some tests fail with journaling. This
> >> > wouldn't take a long time, -j is a temporal thing.
> > We can turn -j into the option to disable journaling after making it
> > the default.
> 
> Then it is better to keep this patch as local patch on your own to
> assist debugging. After we pass all the tests, then -j can be used to
> disable journaling as well as other options.

Do you think the journaling feature can be stabilized in near feature?
If yes, it's okay to me to drop this patch.  Otherwise, we should
accept this change for more developers to be able to test and debug
journaling codes easily (and, IMHO, it's a bit pain to rebase this
patch for each we review journaling codes).

Thanks,

Kazutaka



More information about the sheepdog mailing list