On 02/23, Chao Yu wrote:
> On 2017/2/23 5:55, Jaegeuk Kim wrote:
> > On 02/22, Chao Yu wrote:
> >> On 2017/1/13 6:44, Jaegeuk Kim wrote:
> >>> We don't need to do multiple checkpoints, since we don't actually wait for
> >>> completion of discard commands during checkpoint.
> >>> Instead, we still
On 02/23, Chao Yu wrote:
> On 2017/2/23 5:55, Jaegeuk Kim wrote:
> > On 02/22, Chao Yu wrote:
> >> On 2017/1/13 6:44, Jaegeuk Kim wrote:
> >>> We don't need to do multiple checkpoints, since we don't actually wait for
> >>> completion of discard commands during checkpoint.
> >>> Instead, we still
On 2017/2/23 5:55, Jaegeuk Kim wrote:
> On 02/22, Chao Yu wrote:
>> On 2017/1/13 6:44, Jaegeuk Kim wrote:
>>> We don't need to do multiple checkpoints, since we don't actually wait for
>>> completion of discard commands during checkpoint.
>>> Instead, we still need to avoid very big discard
On 2017/2/23 5:55, Jaegeuk Kim wrote:
> On 02/22, Chao Yu wrote:
>> On 2017/1/13 6:44, Jaegeuk Kim wrote:
>>> We don't need to do multiple checkpoints, since we don't actually wait for
>>> completion of discard commands during checkpoint.
>>> Instead, we still need to avoid very big discard
On 02/22, Chao Yu wrote:
> On 2017/1/13 6:44, Jaegeuk Kim wrote:
> > We don't need to do multiple checkpoints, since we don't actually wait for
> > completion of discard commands during checkpoint.
> > Instead, we still need to avoid very big discard commands, since that large
> > discard can
On 02/22, Chao Yu wrote:
> On 2017/1/13 6:44, Jaegeuk Kim wrote:
> > We don't need to do multiple checkpoints, since we don't actually wait for
> > completion of discard commands during checkpoint.
> > Instead, we still need to avoid very big discard commands, since that large
> > discard can
On 2017/1/13 6:44, Jaegeuk Kim wrote:
> We don't need to do multiple checkpoints, since we don't actually wait for
> completion of discard commands during checkpoint.
> Instead, we still need to avoid very big discard commands, since that large
> discard can interfere block allocation.
I hope we
On 2017/1/13 6:44, Jaegeuk Kim wrote:
> We don't need to do multiple checkpoints, since we don't actually wait for
> completion of discard commands during checkpoint.
> Instead, we still need to avoid very big discard commands, since that large
> discard can interfere block allocation.
I hope we
We don't need to do multiple checkpoints, since we don't actually wait for
completion of discard commands during checkpoint.
Instead, we still need to avoid very big discard commands, since that large
discard can interfere block allocation.
Signed-off-by: Jaegeuk Kim
---
We don't need to do multiple checkpoints, since we don't actually wait for
completion of discard commands during checkpoint.
Instead, we still need to avoid very big discard commands, since that large
discard can interfere block allocation.
Signed-off-by: Jaegeuk Kim
---
10 matches
Mail list logo