Jens Axboe writes:
>>> The problem is really that the WRITE_SYNC is (for Jan's case) behaving
>>> like buffered writes, so it eats up a queue of requests very easily. On
>>> the allocation side, the assumption is that WRITE_SYNC behaves like
>>> dependent reads. Similar to a dd with oflag=direct,
On 2012-12-13 16:02, Jan Kara wrote:
> On Thu 13-12-12 14:30:42, Jens Axboe wrote:
>> On 2012-12-12 20:41, Jeff Moyer wrote:
>>> Jeff Moyer writes:
>>>
> I agree. This isn't about scheduling, we haven't even reached that part
> yet. Back when we split the queues into read vs write, this pr
On Thu 13-12-12 14:30:42, Jens Axboe wrote:
> On 2012-12-12 20:41, Jeff Moyer wrote:
> > Jeff Moyer writes:
> >
> >>> I agree. This isn't about scheduling, we haven't even reached that part
> >>> yet. Back when we split the queues into read vs write, this problem
> >>> obviously wasn't there. Now
Jens Axboe writes:
> On 2012-12-12 20:41, Jeff Moyer wrote:
>> Jeff Moyer writes:
>>
I agree. This isn't about scheduling, we haven't even reached that part
yet. Back when we split the queues into read vs write, this problem
obviously wasn't there. Now we have sync writes and rea
On 2012-12-12 20:41, Jeff Moyer wrote:
> Jeff Moyer writes:
>
>>> I agree. This isn't about scheduling, we haven't even reached that part
>>> yet. Back when we split the queues into read vs write, this problem
>>> obviously wasn't there. Now we have sync writes and reads, both eating
>>> from the
On Wed 12-12-12 14:41:13, Jeff Moyer wrote:
> Jeff Moyer writes:
>
> >> I agree. This isn't about scheduling, we haven't even reached that part
> >> yet. Back when we split the queues into read vs write, this problem
> >> obviously wasn't there. Now we have sync writes and reads, both eating
> >>
On Thu 13-12-12 09:43:31, Shaohua Li wrote:
> 2012/12/12 Jan Kara :
> > On Wed 12-12-12 10:55:15, Shaohua Li wrote:
> >> 2012/12/11 Jan Kara :
> >> > Hi,
> >> >
> >> > I was looking into IO starvation problems where streaming sync writes
> >> > (in
> >> > my case from kjournald but DIO would l
2012/12/12 Jan Kara :
> On Wed 12-12-12 10:55:15, Shaohua Li wrote:
>> 2012/12/11 Jan Kara :
>> > Hi,
>> >
>> > I was looking into IO starvation problems where streaming sync writes (in
>> > my case from kjournald but DIO would look the same) starve reads. This is
>> > because reads happen in s
On Wed, Dec 12, 2012 at 11:26:17AM +0100, Jan Kara wrote:
> On Wed 12-12-12 15:18:21, Dave Chinner wrote:
> > On Wed, Dec 12, 2012 at 03:31:37AM +0100, Jan Kara wrote:
> > > On Tue 11-12-12 16:44:15, Jeff Moyer wrote:
> > > > Jan Kara writes:
> > > >
> > > > > Hi,
> > > > >
> > > > > I was lo
Jeff Moyer writes:
>> I agree. This isn't about scheduling, we haven't even reached that part
>> yet. Back when we split the queues into read vs write, this problem
>> obviously wasn't there. Now we have sync writes and reads, both eating
>> from the same pool. The io scheduler can impact this a
Jens Axboe writes:
> On 2012-12-12 11:11, Jan Kara wrote:
>> On Wed 12-12-12 10:55:15, Shaohua Li wrote:
>>> 2012/12/11 Jan Kara :
Hi,
I was looking into IO starvation problems where streaming sync writes (in
my case from kjournald but DIO would look the same) starve reads
On 2012-12-12 11:11, Jan Kara wrote:
> On Wed 12-12-12 10:55:15, Shaohua Li wrote:
>> 2012/12/11 Jan Kara :
>>> Hi,
>>>
>>> I was looking into IO starvation problems where streaming sync writes (in
>>> my case from kjournald but DIO would look the same) starve reads. This is
>>> because reads h
On Wed 12-12-12 15:18:21, Dave Chinner wrote:
> On Wed, Dec 12, 2012 at 03:31:37AM +0100, Jan Kara wrote:
> > On Tue 11-12-12 16:44:15, Jeff Moyer wrote:
> > > Jan Kara writes:
> > >
> > > > Hi,
> > > >
> > > > I was looking into IO starvation problems where streaming sync writes
> > > > (in
On Wed 12-12-12 10:55:15, Shaohua Li wrote:
> 2012/12/11 Jan Kara :
> > Hi,
> >
> > I was looking into IO starvation problems where streaming sync writes (in
> > my case from kjournald but DIO would look the same) starve reads. This is
> > because reads happen in small chunks and until a reques
On Wed, Dec 12, 2012 at 03:31:37AM +0100, Jan Kara wrote:
> On Tue 11-12-12 16:44:15, Jeff Moyer wrote:
> > Jan Kara writes:
> >
> > > Hi,
> > >
> > > I was looking into IO starvation problems where streaming sync writes
> > > (in
> > > my case from kjournald but DIO would look the same) sta
2012/12/11 Jan Kara :
> Hi,
>
> I was looking into IO starvation problems where streaming sync writes (in
> my case from kjournald but DIO would look the same) starve reads. This is
> because reads happen in small chunks and until a request completes we don't
> start reading further (reader rea
On Tue 11-12-12 16:44:15, Jeff Moyer wrote:
> Jan Kara writes:
>
> > Hi,
> >
> > I was looking into IO starvation problems where streaming sync writes (in
> > my case from kjournald but DIO would look the same) starve reads. This is
> > because reads happen in small chunks and until a request
On Monday 2012-12-10 23:12, Jan Kara wrote:
>
> I was looking into IO starvation problems where streaming sync writes (in
>my case from kjournald but DIO would look the same) starve reads. This is
>because reads happen in small chunks and until a request completes we don't
>start reading further
Jan Kara writes:
> Hi,
>
> I was looking into IO starvation problems where streaming sync writes (in
> my case from kjournald but DIO would look the same) starve reads. This is
> because reads happen in small chunks and until a request completes we don't
> start reading further (reader reads
Hi,
I was looking into IO starvation problems where streaming sync writes (in
my case from kjournald but DIO would look the same) starve reads. This is
because reads happen in small chunks and until a request completes we don't
start reading further (reader reads lots of small files) while wri
20 matches
Mail list logo