[Sheepdog] [PATCH 07/10] sheep: cleanup nr_copies handling

Liu Yuan namei.unix at gmail.com
Mon Apr 30 13:03:12 CEST 2012


On Mon, Apr 30, 2012 at 6:53 PM, Christoph Hellwig <hch at infradead.org> wrote:
> On Mon, Apr 30, 2012 at 04:08:00PM +0900, MORITA Kazutaka wrote:
>> We planned to support a different redundancy level for each VDI.
>> For example:
>>
>>  $ collie vdi create test 1G --copies 1  # no redundancy (for temporary VDIs)
>>  $ collie vdi create test 1G --copies 5  # 5 copies (for important VDIs)
>>  $ collie vdi create test 1G             # use a default redundancy level,
>>                                          # which is specified when formatting cluster
>>
>> inode->nr_copies was the number of copies of the VDI, and
>> sys->nr_copies was the default redundancy level.
>>
>> However, it required a big change in recovery code, so I left it
>> untouched at the time.  For now, I think we should use sys->nr_copies
>> for every places, and make nr_copies in sheepdog_inode and sd_*_req
>> reserved fields.
>
> Makes sense.  I'll submit a patch to do that.
>

I have a pending patch already done with it. [sheep: fix nr_copies in vdi.c]

Thanks,
Yuan



More information about the sheepdog mailing list