Re: [PATCH 0/2] Release 3.2 summary

[Neil Brown <neilb@suse.de>]

On Thu, 27 Jan 2011 17:46:42 +0100
Krzysztof Wojcik <krzysztof.wojcik@intel.com> wrote:

> Neil,
> 
> Intel plans to include following features in mdadm-3.2 release:
> - Array Auto Rebuild,
> - Online Capacity Expansion,
> - Takeover raid10->raid0,
> - Takeover raid0->raid10,
> - Takeover raid1->raid0
> 
> They are almost incorporated into your repository and tested.
> Some of patches are waiting for your acceptation and some will be send
> today.
> We're getting closer to the end of release and I would like to
> summarize and explain what exactly needs to be done.
> 
> To finish our features following patches have to be appyied:
> mdadm:
> - FIX: Meet SET_ARRAY_INFO ioctl requirements (this series)
> - FIX: Remove disks in mdmon for external metadata (this series)
> - WORKAROUND: mdadm hangs during reshape (PART #2)
> - FIX: delta_disk can have UnSet value
> - imsm: FIX: spare cannot be added (very imprtand fix- blocks many
> tests)
> - FIX: start_reshape status should be checked
> - FIX: Container can be left frozen
> - FIX: Array after takeover has to be frozen
> - imsm: FIX: do not allow for container operation for the same disks
> number
> - imsm: Update metadata for second array
> kernel:
> - Add raid1->raid0 takeover support
> - md/raid5: FIX: reshape on degraded devices has wrong configuration
> - md: do not write resync checkpoint, if max_sector has beeen reached
> - unblock the creation of an external metadata RAID if native one
> exists
> 
> Could you review patches above and eventually apply them, please?
> 
> Besides "Assemble: allow to assemble spares on their own" patch has
> to be discused and final solution has to be found.
> 
> After all patches will be applyied we expect all planned featues will
> be working and we can finish the 3.2 release. Of course we will
> continue our testing after official release and new defects may be
> discovered.
> 
> And question for you. When do you plan to announce official 3.2
> release? Can we assume that this will be at the beginning of next
> week?

I think I have applied and pushed out all of the patches in my inbox
that I can agree with or easily fix.  There are a few that raise real
issues but that I am not happy with the approach.  And some that I
don't agree with at all.  But as you say, most have been resolved.

One issue that hasn't been addressed much at all is restarting an array
that is in the middle of a reshape.  I want to at least get that
working for native arrays before I release 3.2.  Getting that working
for external metadata will probably have to wait for 3.2.1.

I very much hope to release 3.2 on Monday.  I will announce it as a
'developer' release - it mostly works but needs lots of testing before
anyone depends on it.

My minimum requirement for 3.2 is that the mdadm test suite passes.

Thanks,
NeilBrown

> 
> ---
> 
> Krzysztof Wojcik (2):
>       FIX: Remove disks in mdmon for external metadata
>       FIX: Meet SET_ARRAY_INFO ioctl requirements
> 
> 
>  Grow.c |   35 +++++++++++++++++++++++------------
>  1 files changed, 23 insertions(+), 12 deletions(-)
>