[sheepdog] [PATCH] sheep: handle recovery request in check_request_in_recovery()
Liu Yuan
namei.unix at gmail.com
Sat Jun 2 16:51:05 CEST 2012
On 06/02/2012 10:46 PM, Christoph Hellwig wrote:
> On Fri, Jun 01, 2012 at 09:10:01PM +0800, Liu Yuan wrote:
>> From: Liu Yuan <tailai.ly at taobao.com>
>>
>> We should call req_done() when is_recoverying_oid() returns true for recovery
>> request, instead of going down to find nothing recovered.
>
> This one causes a regression for me in a test case that first kills one
> zone and then a littler later a second one and then checks that all
> blocks have the expected replication level.
>
> The specific test cases uses 4 initial zones and a replication level of two,
> but I suspect other combinaions would work too.
>
> Figuring out what is going on exactly now.
>
Thanks, this is only changes this series brought to old recovery logic.
This speedup looks valid, so I think other places misbehaves and
uncovered with this patch.
Thanks,
Yuan
More information about the sheepdog
mailing list