[sheepdog] [Sheepdog] [PATCH 0/4] fix a race when multiple sheep join a cluster very quickly

[Yunkai Zhang]

On Wed, May 16, 2012 at 9:21 PM, Christoph Hellwig <hch at infradead.org> wrote:
> On Wed, May 16, 2012 at 09:18:56PM +0800, Liu Yuan wrote:
>> On 05/16/2012 09:13 PM, Christoph Hellwig wrote:
>>
>> > Any chance we can get patches 1-3 in soon as they should be fairly
>> > uncontroverial and easy to review.  I'll probaby revisit patch 4
>> > in coordination with Yunkai Zhang's patch series in this area.
>>
>>
>> Hi Yunkai, is first 3 patches any possible conflict with your unreg/reg
>> group_fd patch? I noticed that you two use different update_epoch_log().
>
> There are various context conflicts, but no functional ones.

Hi Hellwig, Do you have any comments on my latest unregister patch
which have been updated to reflect your previous comments?

Your first 3 patchsets will change these functions:
update_epoch_log()
sd_join_handler()
join()
get_vdi_bitmap_from()
sd_check_join_cb()
...

I think if you have not doubt on this newest unregister patch, I hope
it can be merged before, as this patch is more emergency to sheepdog.
We need it to speed up sheepdop when processing join/leave/notify
events.



>



-- 
Yunkai Zhang
Work at Taobao