mimic-dev1 branch for 13.0.1

mimic-dev1 branch for 13.0.1

[Sage Weil]

Hi all,

I've pushed a mimic-dev1 branch and am kicking off a bunch of qa runs for 
the 13.0.1 dev checkpoint release.  The goal here is to make sure these QA 
runs are clean so that we don't have weeks and weeks of cleaning up our 
testing debt at the end of the mimic cycle.  Yuri is on PTO right now so 
it'll be up to the leads to drive these tests for now.

Overall for the rados suite I think we are in reasonably good shape, 
but there are definitely some failures that we should clean up before 
moving on.

Please try to merge fixes directly into the mimic-dev1 branch (instead of 
cherry-picking the fixes).  We will periodically merge mimic-dev1 back 
into master, and also merge the eventual 13.0.1 release tag into master as 
well.

Thanks!
sage

Re: mimic-dev1 branch for 13.0.1

[kefu chai]

On Tue, Dec 5, 2017 at 1:57 AM, Sage Weil <sweil@redhat.com> wrote:
> Hi all,
>
> I've pushed a mimic-dev1 branch and am kicking off a bunch of qa runs for
> the 13.0.1 dev checkpoint release.  The goal here is to make sure these QA
> runs are clean so that we don't have weeks and weeks of cleaning up our
> testing debt at the end of the mimic cycle.  Yuri is on PTO right now so
> it'll be up to the leads to drive these tests for now.
>
> Overall for the rados suite I think we are in reasonably good shape,
> but there are definitely some failures that we should clean up before
> moving on.
>
> Please try to merge fixes directly into the mimic-dev1 branch (instead of
> cherry-picking the fixes).  We will periodically merge mimic-dev1 back
> into master, and also merge the eventual 13.0.1 release tag into master as
> well.

some fixes also address the QA suite failures on master branch, if they go to
mimic-dev1 and will not go back to master branch immediately. we will still be
seeing test failures in the merge window. take
https://github.com/ceph/ceph/pull/19395,
as an example, it is in mimic-dev1 first. but i found that the same failure also
existed in master, that's why i want to have this fix in master before waiting
the merge in batch.

so, shall we just be more patient in this case, or we can forward-port
the fix to
master immediately?

-- 
Regards
Kefu Chai