On 11/30/2017 07:38 PM, 彭良彦 wrote:
> On 17/12/1 上午10:38, Michael Lyle wrote:
>> On 11/30/2017 06:30 PM, 彭良彦 wrote:
>>> If the cache device is not persistent like ramdisk, all data will be lost after reboot. We have to use make-bcache to fix the inconsistent super block of backing device.
>>
>> Using a non-persistent writeback cache is guaranteed to lose data.
>> (Writes will only go to the cache device and not make it to the
>> persistent volume).
>>
>> Mike
>
> If we use ramdisk as a writethrough cache to improve the read performance, the backing device is a network disk, the super block state of backing device still has chance to be inconsistent due to disconnected network. Like in above reproduce steps, i didn't use a cache device. :-)N�����r��y���b�X��ǧv�^�){.n�+����{�mƜ�觶��ܨ}���Ơz�&j:+v�������zZ+��+zf���h���~����i���z� �w���?����&�)ߢfl===
>
I cannot accept this patch. It is guaranteed to corrupt data, by
starting the client device when the cache is in a bad state.
If you use 'writeback', you must have a cache device and the cache
device must be present to start the bcache client device.
Mike
--
To unsubscribe from this list: send the line "unsubscribe linux-bcache" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
