Hi Jens,
I'm chasing a performance bottleneck identified by tiobench that seems
to be caused by CFQ. On a SLES10-SP3 kernel (2.6.16, with some patches
moving cfq closer to 2.6.17) tiobench with 8 threads gets about 260MB/s
sequential read throughput. On a recent kernels (including vanilla
2.6.34-rc) it makes about 145MB/s, a regression of 45%. The queue and
readahead parameters are the same.
This goes back some time, 2.6.27 already seems to have a bad
performance.
Changing the scheduler to noop will increase the throughput back into
the 260MB/s range. So this is not a driver issue.
Also increasing quantum *and* readahead will increase the throughput,
but not by as much. Both noop and these tweaks decrease the write
throughput somewhat however...
Apparently on recent kernels the number of dispatched requests stays
mostly at or below 4 and the dispatched sector count at or below 2000,
which is not enough to fill the bandwidth on this setup.
On 2.6.16 the number of dispatched requests hovers around 22 and the
sector count around 16000.
I uploaded blktraces for the read part of the tiobench runs for both
2.6.16 and 2.6.32:
http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/
Do you have any idea about the cause of this regression?
Thanks,
Miklos
On Fri, Apr 16, 2010 at 02:27:58PM +0200, Miklos Szeredi wrote:
> Hi Jens,
>
> I'm chasing a performance bottleneck identified by tiobench that seems
> to be caused by CFQ. On a SLES10-SP3 kernel (2.6.16, with some patches
> moving cfq closer to 2.6.17) tiobench with 8 threads gets about 260MB/s
> sequential read throughput. On a recent kernels (including vanilla
> 2.6.34-rc) it makes about 145MB/s, a regression of 45%. The queue and
> readahead parameters are the same.
I've also just noticed this using the most recent Redhat kernels. Writes don't
seem to be affected at all. If the latest Redhat kernels mean anything here, I
might as well show you what I've got, in case there is something common.
./disktest -B 4k -h 1 -I BD -K 32 -p l -P T -T 300 -r /dev/sdf
With cfq we get this:
STAT | 17260 | v1.4.2 | /dev/sdf | Heartbeat read throughput: 15032320.0B/s (14.34MB/s), IOPS 3670.0/s
And with noop we get this:
STAT | 17260 | v1.4.2 | /dev/sdf | Heartbeat read throughput: 111759360.0B/s (106.58MB/s), IOPS 27285.0/s.
Setting some very large and busy web servers to noop just out of curiousity
also reduced the average io time and dropped the load.
Chris
Hi Miklos,
I don't think this is related to CFQ. I've made a graph of the
accessed (read) sectors (see attached).
You can see that the green cloud (2.6.16) is much more concentrated,
while the red one (2.6.32) is split in two, and you can better
recognize the different lines.
This means that the FS put more distance between the blocks of the
files written by the tio threads, and the read time is therefore
impacted, since the disk head has to perform longer seeks. On the
other hand, if you read those files sequentially with a single thread,
the performance may be better with the new layout, so YMMV. When
testing 2.6.32 and up, you should consider testing also with
low_latency setting disabled, since tuning for latency can negatively
affect throughput.
Thanks,
Corrado
On Fri, Apr 16, 2010 at 2:27 PM, Miklos Szeredi <[email protected]> wrote:
> Hi Jens,
>
> I'm chasing a performance bottleneck identified by tiobench that seems
> to be caused by CFQ. On a SLES10-SP3 kernel (2.6.16, with some patches
> moving cfq closer to 2.6.17) tiobench with 8 threads gets about 260MB/s
> sequential read throughput. On a recent kernels (including vanilla
> 2.6.34-rc) it makes about 145MB/s, a regression of 45%. The queue and
> readahead parameters are the same.
>
> This goes back some time, 2.6.27 already seems to have a bad
> performance.
>
> Changing the scheduler to noop will increase the throughput back into
> the 260MB/s range. So this is not a driver issue.
>
> Also increasing quantum *and* readahead will increase the throughput,
> but not by as much. Both noop and these tweaks decrease the write
> throughput somewhat however...
>
> Apparently on recent kernels the number of dispatched requests stays
> mostly at or below 4 and the dispatched sector count at or below 2000,
> which is not enough to fill the bandwidth on this setup.
>
> On 2.6.16 the number of dispatched requests hovers around 22 and the
> sector count around 16000.
>
> I uploaded blktraces for the read part of the tiobench runs for both
> 2.6.16 and 2.6.32:
>
> http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/
>
> Do you have any idea about the cause of this regression?
>
> Thanks,
> Miklos
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
--
__________________________________________________________________________
dott. Corrado Zoccolo mailto:[email protected]
PhD - Department of Computer Science - University of Pisa, Italy
--------------------------------------------------------------------------
The self-confidence of a warrior is not the self-confidence of the average
man. The average man seeks certainty in the eyes of the onlooker and calls
that self-confidence. The warrior seeks impeccability in his own eyes and
calls that humbleness.
Tales of Power - C. Castaneda
Hi Corrado,
On Sat, 2010-04-17 at 14:46 +0200, Corrado Zoccolo wrote:
> I don't think this is related to CFQ. I've made a graph of the
> accessed (read) sectors (see attached).
> You can see that the green cloud (2.6.16) is much more concentrated,
> while the red one (2.6.32) is split in two, and you can better
> recognize the different lines.
> This means that the FS put more distance between the blocks of the
> files written by the tio threads, and the read time is therefore
> impacted, since the disk head has to perform longer seeks. On the
> other hand, if you read those files sequentially with a single thread,
> the performance may be better with the new layout, so YMMV. When
> testing 2.6.32 and up, you should consider testing also with
> low_latency setting disabled, since tuning for latency can negatively
> affect throughput.
low_latency is set to zero in all tests.
The layout difference doesn't explain why setting the scheduler to
"noop" consistently speeds up read throughput in 8-thread tiobench to
almost twice. This fact alone pretty clearly indicates that the I/O
scheduler is the culprit.
There are other indications, see the attached btt output for both
traces. From there it appears that 2.6.16 does more and longer seeks,
yet it's getting an overall better performance.
I've also tested with plain "dd" instead of tiobench where the
filesystem layout stayed exactly the same between tests. Still the
speed difference is there.
Thanks,
Miklos
************************************************************************
btt output for 2.6.16:
==================== All Devices ====================
ALL MIN AVG MAX N
--------------- ------------- ------------- ------------- -----------
Q2Q 0.000000047 0.000854417 1.003550405 67465
Q2G 0.000000458 0.000001211 0.000123527 46062
G2I 0.000000123 0.000001815 0.000494517 46074
Q2M 0.000000186 0.000001798 0.000010296 21404
I2D 0.000000162 0.000158001 0.040794333 46062
M2D 0.000000878 0.000133130 0.040585566 21404
D2C 0.000053870 0.023778266 0.234154543 67466
Q2C 0.000056746 0.023931014 0.234176000 67466
==================== Device Overhead ====================
DEV | Q2G G2I Q2M I2D D2C
---------- | --------- --------- --------- --------- ---------
( 8, 64) | 0.0035% 0.0052% 0.0024% 0.4508% 99.3617%
---------- | --------- --------- --------- --------- ---------
Overall | 0.0035% 0.0052% 0.0024% 0.4508% 99.3617%
==================== Device Merge Information ====================
DEV | #Q #D Ratio | BLKmin BLKavg BLKmax Total
---------- | -------- -------- ------- | -------- -------- -------- --------
( 8, 64) | 67466 46062 1.5 | 8 597 1024 27543688
==================== Device Q2Q Seek Information ====================
DEV | NSEEKS MEAN MEDIAN | MODE
---------- | --------------- --------------- --------------- | ---------------
( 8, 64) | 67466 866834.0 0 | 0(34558)
---------- | --------------- --------------- --------------- | ---------------
Overall | NSEEKS MEAN MEDIAN | MODE
Average | 67466 866834.0 0 | 0(34558)
==================== Device D2D Seek Information ====================
DEV | NSEEKS MEAN MEDIAN | MODE
---------- | --------------- --------------- --------------- | ---------------
( 8, 64) | 46062 1265503.9 0 | 0(13242)
---------- | --------------- --------------- --------------- | ---------------
Overall | NSEEKS MEAN MEDIAN | MODE
Average | 46062 1265503.9 0 | 0(13242)
==================== Plug Information ====================
DEV | # Plugs # Timer Us | % Time Q Plugged
---------- | ---------- ---------- | ----------------
( 8, 64) | 29271( 533) | 3.878105328%
DEV | IOs/Unp IOs/Unp(to)
---------- | ---------- ----------
( 8, 64) | 19.2 19.7
---------- | ---------- ----------
Overall | IOs/Unp IOs/Unp(to)
Average | 19.2 19.7
==================== Active Requests At Q Information ====================
DEV | Avg Reqs @ Q
---------- | -------------
( 8, 64) | 0.8
==================== I/O Active Period Information ====================
DEV | # Live Avg. Act Avg. !Act % Live
---------- | ---------- ------------- ------------- ------
( 8, 64) | 545 0.100012237 0.005766640 94.56
---------- | ---------- ------------- ------------- ------
Total Sys | 545 0.100012237 0.005766640 94.56
************************************************************************
btt output for 2.6.32:
==================== All Devices ====================
ALL MIN AVG MAX N
--------------- ------------- ------------- ------------- -----------
Q2Q 0.000000279 0.001710581 1.803934429 69429
Q2G 0.000000908 0.001798735 23.144764798 54940
S2G 0.022460311 6.581680621 23.144763751 15
G2I 0.000000628 0.000001576 0.000120409 54942
Q2M 0.000000628 0.000001611 0.000013201 14490
I2D 0.000000768 0.289812035 86.820205789 54940
M2D 0.000005518 0.098208187 0.794441158 14490
D2C 0.000173141 0.008056256 0.219516385 69430
Q2C 0.000179077 0.259305605 86.820559403 69430
==================== Device Overhead ====================
DEV | Q2G G2I Q2M I2D D2C
---------- | --------- --------- --------- --------- ---------
( 8, 64) | 0.5489% 0.0005% 0.0001% 88.4394% 3.1069%
---------- | --------- --------- --------- --------- ---------
Overall | 0.5489% 0.0005% 0.0001% 88.4394% 3.1069%
==================== Device Merge Information ====================
DEV | #Q #D Ratio | BLKmin BLKavg BLKmax Total
---------- | -------- -------- ------- | -------- -------- -------- --------
( 8, 64) | 69430 54955 1.3 | 8 520 2048 28614984
==================== Device Q2Q Seek Information ====================
DEV | NSEEKS MEAN MEDIAN | MODE
---------- | --------------- --------------- --------------- | ---------------
( 8, 64) | 69430 546377.3 0 | 0(50235)
---------- | --------------- --------------- --------------- | ---------------
Overall | NSEEKS MEAN MEDIAN | MODE
Average | 69430 546377.3 0 | 0(50235)
==================== Device D2D Seek Information ====================
DEV | NSEEKS MEAN MEDIAN | MODE
---------- | --------------- --------------- --------------- | ---------------
( 8, 64) | 54955 565286.3 0 | 0(37535)
---------- | --------------- --------------- --------------- | ---------------
Overall | NSEEKS MEAN MEDIAN | MODE
Average | 54955 565286.3 0 | 0(37535)
==================== Plug Information ====================
DEV | # Plugs # Timer Us | % Time Q Plugged
---------- | ---------- ---------- | ----------------
( 8, 64) | 2310( 0) | 0.049353257%
DEV | IOs/Unp IOs/Unp(to)
---------- | ---------- ----------
( 8, 64) | 7.3 0.0
---------- | ---------- ----------
Overall | IOs/Unp IOs/Unp(to)
Average | 7.3 0.0
==================== Active Requests At Q Information ====================
DEV | Avg Reqs @ Q
---------- | -------------
( 8, 64) | 132.8
==================== I/O Active Period Information ====================
DEV | # Live Avg. Act Avg. !Act % Live
---------- | ---------- ------------- ------------- ------
( 8, 64) | 4835 0.023848998 0.000714665 97.09
---------- | ---------- ------------- ------------- ------
Total Sys | 4835 0.023848998 0.000714665 97.09
On Mon, Apr 19, 2010 at 1:46 PM, Miklos Szeredi <[email protected]> wrote:> Hi Corrado,>> On Sat, 2010-04-17 at 14:46 +0200, Corrado Zoccolo wrote:>> I don't think this is related to CFQ. I've made a graph of the>> accessed (read) sectors (see attached).>> You can see that the green cloud (2.6.16) is much more concentrated,>> while the red one (2.6.32) is split in two, and you can better>> recognize the different lines.>> This means that the FS put more distance between the blocks of the>> files written by the tio threads, and the read time is therefore>> impacted, since the disk head has to perform longer seeks. On the>> other hand, if you read those files sequentially with a single thread,>> the performance may be better with the new layout, so YMMV. When>> testing 2.6.32 and up, you should consider testing also with>> low_latency setting disabled, since tuning for latency can negatively>> affect throughput.Hi Miklos,can you give more information about the setup?How much memory do you have, what is the disk configuration (is this ahw raid?) and so on.>> low_latency is set to zero in all tests.>> The layout difference doesn't explain why setting the scheduler to> "noop" consistently speeds up read throughput in 8-thread tiobench to> almost twice. This fact alone pretty clearly indicates that the I/O> scheduler is the culprit.From the attached btt output, I see that a lot of time is spentwaiting to allocate new request structures.> S2G 0.022460311 6.581680621 23.144763751 15Since noop doesn't attach fancy data to each request, it can savethose allocations, thus resulting in no sleeps.The delays in allocation, though, may not be completely imputable tothe I/O scheduler, and working in constrained memory conditions willnegatively affect it.
> There are other indications, see the attached btt output for both> traces. From there it appears that 2.6.16 does more and longer seeks,> yet it's getting an overall better performance.I see less seeks for 2.6.16, but longer on average.It seems that 2.6.16 allows more requests from the same process to bestreamed to disk before switching to an other process.Since the timeslice is the same, it might be that we are limiting thenumber of requests per queue due to memory congestion.
> I've also tested with plain "dd" instead of tiobench where the> filesystem layout stayed exactly the same between tests. Still the> speed difference is there.Does dropping caches before the read test change the situation?
Thanks,Corrado>> Thanks,> Miklos>> ************************************************************************> btt output for 2.6.16:> ==================== All Devices ====================>> ALL MIN AVG MAX N> --------------- ------------- ------------- ------------- ----------->> Q2Q 0.000000047 0.000854417 1.003550405 67465> Q2G 0.000000458 0.000001211 0.000123527 46062> G2I 0.000000123 0.000001815 0.000494517 46074> Q2M 0.000000186 0.000001798 0.000010296 21404> I2D 0.000000162 0.000158001 0.040794333 46062> M2D 0.000000878 0.000133130 0.040585566 21404> D2C 0.000053870 0.023778266 0.234154543 67466> Q2C 0.000056746 0.023931014 0.234176000 67466>> ==================== Device Overhead ====================>> DEV | Q2G G2I Q2M I2D D2C> ---------- | --------- --------- --------- --------- ---------> ( 8, 64) | 0.0035% 0.0052% 0.0024% 0.4508% 99.3617%> ---------- | --------- --------- --------- --------- ---------> Overall | 0.0035% 0.0052% 0.0024% 0.4508% 99.3617%>> ==================== Device Merge Information ====================>> DEV | #Q #D Ratio | BLKmin BLKavg BLKmax Total> ---------- | -------- -------- ------- | -------- -------- -------- --------> ( 8, 64) | 67466 46062 1.5 | 8 597 1024 27543688>> ==================== Device Q2Q Seek Information ====================>> DEV | NSEEKS MEAN MEDIAN | MODE> ---------- | --------------- --------------- --------------- | ---------------> ( 8, 64) | 67466 866834.0 0 | 0(34558)> ---------- | --------------- --------------- --------------- | ---------------> Overall | NSEEKS MEAN MEDIAN | MODE> Average | 67466 866834.0 0 | 0(34558)>> ==================== Device D2D Seek Information ====================>> DEV | NSEEKS MEAN MEDIAN | MODE> ---------- | --------------- --------------- --------------- | ---------------> ( 8, 64) | 46062 1265503.9 0 | 0(13242)> ---------- | --------------- --------------- --------------- | ---------------> Overall | NSEEKS MEAN MEDIAN | MODE> Average | 46062 1265503.9 0 | 0(13242)>> ==================== Plug Information ====================>> DEV | # Plugs # Timer Us | % Time Q Plugged> ---------- | ---------- ---------- | ----------------> ( 8, 64) | 29271( 533) | 3.878105328%>> DEV | IOs/Unp IOs/Unp(to)> ---------- | ---------- ----------> ( 8, 64) | 19.2 19.7> ---------- | ---------- ----------> Overall | IOs/Unp IOs/Unp(to)> Average | 19.2 19.7>> ==================== Active Requests At Q Information ====================>> DEV | Avg Reqs @ Q> ---------- | -------------> ( 8, 64) | 0.8>> ==================== I/O Active Period Information ====================>> DEV | # Live Avg. Act Avg. !Act % Live> ---------- | ---------- ------------- ------------- ------> ( 8, 64) | 545 0.100012237 0.005766640 94.56> ---------- | ---------- ------------- ------------- ------> Total Sys | 545 0.100012237 0.005766640 94.56>> ************************************************************************> btt output for 2.6.32:>> ==================== All Devices ====================>> ALL MIN AVG MAX N> --------------- ------------- ------------- ------------- ----------->> Q2Q 0.000000279 0.001710581 1.803934429 69429> Q2G 0.000000908 0.001798735 23.144764798 54940> S2G 0.022460311 6.581680621 23.144763751 15> G2I 0.000000628 0.000001576 0.000120409 54942> Q2M 0.000000628 0.000001611 0.000013201 14490> I2D 0.000000768 0.289812035 86.820205789 54940> M2D 0.000005518 0.098208187 0.794441158 14490> D2C 0.000173141 0.008056256 0.219516385 69430> Q2C 0.000179077 0.259305605 86.820559403 69430>> ==================== Device Overhead ====================>> DEV | Q2G G2I Q2M I2D D2C> ---------- | --------- --------- --------- --------- ---------> ( 8, 64) | 0.5489% 0.0005% 0.0001% 88.4394% 3.1069%> ---------- | --------- --------- --------- --------- ---------> Overall | 0.5489% 0.0005% 0.0001% 88.4394% 3.1069%>> ==================== Device Merge Information ====================>> DEV | #Q #D Ratio | BLKmin BLKavg BLKmax Total> ---------- | -------- -------- ------- | -------- -------- -------- --------> ( 8, 64) | 69430 54955 1.3 | 8 520 2048 28614984>> ==================== Device Q2Q Seek Information ====================>> DEV | NSEEKS MEAN MEDIAN | MODE> ---------- | --------------- --------------- --------------- | ---------------> ( 8, 64) | 69430 546377.3 0 | 0(50235)> ---------- | --------------- --------------- --------------- | ---------------> Overall | NSEEKS MEAN MEDIAN | MODE> Average | 69430 546377.3 0 | 0(50235)>> ==================== Device D2D Seek Information ====================>> DEV | NSEEKS MEAN MEDIAN | MODE> ---------- | --------------- --------------- --------------- | ---------------> ( 8, 64) | 54955 565286.3 0 | 0(37535)> ---------- | --------------- --------------- --------------- | ---------------> Overall | NSEEKS MEAN MEDIAN | MODE> Average | 54955 565286.3 0 | 0(37535)>> ==================== Plug Information ====================>> DEV | # Plugs # Timer Us | % Time Q Plugged> ---------- | ---------- ---------- | ----------------> ( 8, 64) | 2310( 0) | 0.049353257%>> DEV | IOs/Unp IOs/Unp(to)> ---------- | ---------- ----------> ( 8, 64) | 7.3 0.0> ---------- | ---------- ----------> Overall | IOs/Unp IOs/Unp(to)> Average | 7.3 0.0>> ==================== Active Requests At Q Information ====================>> DEV | Avg Reqs @ Q> ---------- | -------------> ( 8, 64) | 132.8>> ==================== I/O Active Period Information ====================>> DEV | # Live Avg. Act Avg. !Act % Live> ---------- | ---------- ------------- ------------- ------> ( 8, 64) | 4835 0.023848998 0.000714665 97.09> ---------- | ---------- ------------- ------------- ------> Total Sys | 4835 0.023848998 0.000714665 97.09>>>>????{.n?+???????+%?????ݶ??w??{.n?+????{??G?????{ay? ʇڙ?,j??f???h?????????z_??(?階?ݢj"???m??????G????????????&???~???iO???z??v?^?m???? ????????I?
Corrado,
On Tue, 2010-04-20 at 22:50 +0200, Corrado Zoccolo wrote:
> can you give more information about the setup?
> How much memory do you have, what is the disk configuration (is this a
> hw raid?) and so on.
8G of memory 8-way Xeon CPU, fiber channel attached storage array (HP
HSV200). I don't know the configuration of the array.
> > low_latency is set to zero in all tests.
> >
> > The layout difference doesn't explain why setting the scheduler to
> > "noop" consistently speeds up read throughput in 8-thread tiobench to
> > almost twice. This fact alone pretty clearly indicates that the I/O
> > scheduler is the culprit.
> From the attached btt output, I see that a lot of time is spent
> waiting to allocate new request structures.
> > S2G 0.022460311 6.581680621 23.144763751 15
> Since noop doesn't attach fancy data to each request, it can save
> those allocations, thus resulting in no sleeps.
> The delays in allocation, though, may not be completely imputable to
> the I/O scheduler, and working in constrained memory conditions will
> negatively affect it.
I verified with the simple dd test that this happens even if there's no
memory pressure from the cache by dd-ing only 5G of files, after
clearing the cache. This way ~2G of memory are completely free
throughout the test.
> > I've also tested with plain "dd" instead of tiobench where the
> > filesystem layout stayed exactly the same between tests. Still the
> > speed difference is there.
> Does dropping caches before the read test change the situation?
In all my tests I drop caches before running it.
Please let me know if you need more information.
Thanks,
Miklos
Jens, Corrado,
Here's a graph showing the number of issued but not yet completed
requests versus time for CFQ and NOOP schedulers running the tiobench
benchmark with 8 threads:
http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/queue-depth.jpg
It shows pretty clearly the performance problem is because CFQ is not
issuing enough request to fill the bandwidth.
Is this the correct behavior of CFQ or is this a bug?
This is on a vanilla 2.6.34-rc4 kernel with two tunables modified:
read_ahead_kb=512
low_latency=0 (for CFQ)
Thanks,
Miklos
Hi Miklos,
On Wed, Apr 21, 2010 at 6:05 PM, Miklos Szeredi <[email protected]> wrote:
> Jens, Corrado,
>
> Here's a graph showing the number of issued but not yet completed
> requests versus time for CFQ and NOOP schedulers running the tiobench
> benchmark with 8 threads:
>
> http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/queue-depth.jpg
>
> It shows pretty clearly the performance problem is because CFQ is not
> issuing enough request to fill the bandwidth.
>
> Is this the correct behavior of CFQ or is this a bug?
This is the expected behavior from CFQ, even if it is not optimal,
since we aren't able to identify multi-splindle disks yet. Can you
post the result of "grep -r . ." in your /sys/block/*/queue directory,
to see if we can find some parameter that can help identifying your
hardware as a multi-spindle disk.
>
> This is on a vanilla 2.6.34-rc4 kernel with two tunables modified:
>
> read_ahead_kb=512
> low_latency=0 (for CFQ)
You should get much better throughput by setting
/sys/block/_your_disk_/queue/iosched/slice_idle to 0, or
/sys/block/_your_disk_/queue/rotational to 0.
Thanks,
Corrado
>
> Thanks,
> Miklos
>
>
>
On Thu, 2010-04-22 at 09:59 +0200, Corrado Zoccolo wrote:
> Hi Miklos,
> On Wed, Apr 21, 2010 at 6:05 PM, Miklos Szeredi <[email protected]> wrote:
> > Jens, Corrado,
> >
> > Here's a graph showing the number of issued but not yet completed
> > requests versus time for CFQ and NOOP schedulers running the tiobench
> > benchmark with 8 threads:
> >
> > http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/queue-depth.jpg
> >
> > It shows pretty clearly the performance problem is because CFQ is not
> > issuing enough request to fill the bandwidth.
> >
> > Is this the correct behavior of CFQ or is this a bug?
> This is the expected behavior from CFQ, even if it is not optimal,
> since we aren't able to identify multi-splindle disks yet. Can you
> post the result of "grep -r . ." in your /sys/block/*/queue directory,
> to see if we can find some parameter that can help identifying your
> hardware as a multi-spindle disk.
./iosched/quantum:8
./iosched/fifo_expire_sync:124
./iosched/fifo_expire_async:248
./iosched/back_seek_max:16384
./iosched/back_seek_penalty:2
./iosched/slice_sync:100
./iosched/slice_async:40
./iosched/slice_async_rq:2
./iosched/slice_idle:8
./iosched/low_latency:0
./iosched/group_isolation:0
./nr_requests:128
./read_ahead_kb:512
./max_hw_sectors_kb:32767
./max_sectors_kb:512
./max_segments:64
./max_segment_size:65536
./scheduler:noop deadline [cfq]
./hw_sector_size:512
./logical_block_size:512
./physical_block_size:512
./minimum_io_size:512
./optimal_io_size:0
./discard_granularity:0
./discard_max_bytes:0
./discard_zeroes_data:0
./rotational:1
./nomerges:0
./rq_affinity:1
> >
> > This is on a vanilla 2.6.34-rc4 kernel with two tunables modified:
> >
> > read_ahead_kb=512
> > low_latency=0 (for CFQ)
> You should get much better throughput by setting
> /sys/block/_your_disk_/queue/iosched/slice_idle to 0, or
> /sys/block/_your_disk_/queue/rotational to 0.
slice_idle=0 definitely helps. rotational=0 seems to help on 2.6.34-rc
but not on 2.6.32.
As far as I understand setting slice_idle to zero is just a workaround
to make cfq look at all the other queues instead of serving one
exclusively for a long time.
I have very little understanding of I/O scheduling but my idea of what's
really needed here is to realize that one queue is not able to saturate
the device and there's a large backlog of requests on other queues that
are waiting to be served. Is something like that implementable?
Thanks,
Miklos
On Thu 22-04-10 12:23:29, Miklos Szeredi wrote:
> > >
> > > This is on a vanilla 2.6.34-rc4 kernel with two tunables modified:
> > >
> > > read_ahead_kb=512
> > > low_latency=0 (for CFQ)
> > You should get much better throughput by setting
> > /sys/block/_your_disk_/queue/iosched/slice_idle to 0, or
> > /sys/block/_your_disk_/queue/rotational to 0.
>
> slice_idle=0 definitely helps. rotational=0 seems to help on 2.6.34-rc
> but not on 2.6.32.
>
> As far as I understand setting slice_idle to zero is just a workaround
> to make cfq look at all the other queues instead of serving one
> exclusively for a long time.
Yes, basically it disables idling (i.e., waiting whether a thread sends
more IO so that we can get better IO locality).
> I have very little understanding of I/O scheduling but my idea of what's
> really needed here is to realize that one queue is not able to saturate
> the device and there's a large backlog of requests on other queues that
> are waiting to be served. Is something like that implementable?
I see a problem with defining "saturate the device" - but maybe we could
measure something like "completed requests / sec" and try autotuning
slice_idle to maximize this value (hopefully the utility function should
be concave so we can just use "local optimization").
Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR
On Thu, Apr 22, 2010 at 09:59:14AM +0200, Corrado Zoccolo wrote:
> Hi Miklos,
> On Wed, Apr 21, 2010 at 6:05 PM, Miklos Szeredi <[email protected]> wrote:
> > Jens, Corrado,
> >
> > Here's a graph showing the number of issued but not yet completed
> > requests versus time for CFQ and NOOP schedulers running the tiobench
> > benchmark with 8 threads:
> >
> > http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/queue-depth.jpg
> >
> > It shows pretty clearly the performance problem is because CFQ is not
> > issuing enough request to fill the bandwidth.
> >
> > Is this the correct behavior of CFQ or is this a bug?
> This is the expected behavior from CFQ, even if it is not optimal,
> since we aren't able to identify multi-splindle disks yet.
In the past we were of the opinion that for sequential workload multi spindle
disks will not matter much as readahead logic (in OS and possibly in
hardware also) will help. For random workload we anyway don't idle on the
single cfqq so it is fine. But my tests now seem to be telling a different
story.
I also have one FC link to one of the HP EVA and I am running increasing
number of sequential readers to see if throughput goes up as number of
readers go up. The results are with noop and cfq. I do flush OS caches
across the runs but I have no control on caching on HP EVA.
Kernel=2.6.34-rc5
DIR=/mnt/iostestmnt/fio DEV=/dev/mapper/mpathe
Workload=bsr iosched=cfq Filesz=2G bs=4K
=========================================================================
job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
--- --- -- ------------ ----------- ------------- -----------
bsr 1 1 135366 59024 0 0
bsr 1 2 124256 126808 0 0
bsr 1 4 132921 341436 0 0
bsr 1 8 129807 392904 0 0
bsr 1 16 129988 773991 0 0
Kernel=2.6.34-rc5
DIR=/mnt/iostestmnt/fio DEV=/dev/mapper/mpathe
Workload=bsr iosched=noop Filesz=2G bs=4K
=========================================================================
job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
--- --- -- ------------ ----------- ------------- -----------
bsr 1 1 126187 95272 0 0
bsr 1 2 185154 72908 0 0
bsr 1 4 224622 88037 0 0
bsr 1 8 285416 115592 0 0
bsr 1 16 348564 156846 0 0
So in case of NOOP, throughput shotup to 348MB/s but CFQ reamains more or
less constat, about 130MB/s.
So atleast in this case, a single sequential CFQ queue is not keeing the
disk busy enough.
I am wondering why my testing results were different in the past. May be
it was a different piece of hardware and behavior various across hardware?
Anyway, if that's the case, then we probably need to allow IO from
multiple sequential readers and keep a watch on throughput. If throughput
drops then reduce the number of parallel sequential readers. Not sure how
much of code that is but with multiple cfqq going in parallel, ioprio
logic will more or less stop working in CFQ (on multi-spindle hardware).
FWIW, I also ran tiobench on same HP EVA with NOOP and CFQ. And indeed
Read throughput is bad with CFQ.
With NOOP
=========
# /usr/bin/tiotest -t 8 -f 2000 -r 4000 -b 4096 -d /mnt/mpathe
Tiotest results for 8 concurrent io threads:
,----------------------------------------------------------------------.
| Item | Time | Rate | Usr CPU | Sys CPU |
+-----------------------+----------+--------------+----------+---------+
| Write 16000 MBs | 44.1 s | 362.410 MB/s | 25.3 % | 1239.4 % |
| Random Write 125 MBs | 0.8 s | 156.182 MB/s | 19.7 % | 484.8 % |
| Read 16000 MBs | 59.9 s | 267.008 MB/s | 12.4 % | 197.1 % |
| Random Read 125 MBs | 16.7 s | 7.478 MB/s | 1.0 % | 23.7 % |
`----------------------------------------------------------------------'
Tiotest latency results:
,-------------------------------------------------------------------------.
| Item | Average latency | Maximum latency | % >2 sec | % >10 sec |
+--------------+-----------------+-----------------+----------+-----------+
| Write | 0.083 ms | 834.092 ms | 0.00000 | 0.00000 |
| Random Write | 0.021 ms | 21.024 ms | 0.00000 | 0.00000 |
| Read | 0.115 ms | 105.830 ms | 0.00000 | 0.00000 |
| Random Read | 4.088 ms | 295.605 ms | 0.00000 | 0.00000 |
|--------------+-----------------+-----------------+----------+-----------|
| Total | 0.114 ms | 834.092 ms | 0.00000 | 0.00000 |
`--------------+-----------------+-----------------+----------+-----------'
With CFQ
========
# /usr/bin/tiotest -t 8 -f 2000 -r 4000 -b 4096 -d /mnt/mpathe
Tiotest results for 8 concurrent io threads:
,----------------------------------------------------------------------.
| Item | Time | Rate | Usr CPU | Sys CPU |
+-----------------------+----------+--------------+----------+---------+
| Write 16000 MBs | 49.5 s | 323.086 MB/s | 21.7 % | 1175.6 % |
| Random Write 125 MBs | 2.2 s | 57.148 MB/s | 5.0 % | 188.1 % |
| Read 16000 MBs | 162.7 s | 98.311 MB/s | 4.7 % | 71.0 % |
| Random Read 125 MBs | 17.0 s | 7.344 MB/s | 0.8 % | 26.5 % |
`----------------------------------------------------------------------'
Tiotest latency results:
,-------------------------------------------------------------------------.
| Item | Average latency | Maximum latency | % >2 sec | % >10 sec |
+--------------+-----------------+-----------------+----------+-----------+
| Write | 0.093 ms | 832.680 ms | 0.00000 | 0.00000 |
| Random Write | 0.017 ms | 12.031 ms | 0.00000 | 0.00000 |
| Read | 0.316 ms | 561.623 ms | 0.00000 | 0.00000 |
| Random Read | 4.126 ms | 273.156 ms | 0.00000 | 0.00000 |
|--------------+-----------------+-----------------+----------+-----------|
| Total | 0.219 ms | 832.680 ms | 0.00000 | 0.00000 |
`--------------+-----------------+-----------------+----------+-----------'
Thanks
Vivek
On Thu, 2010-04-22 at 17:53 +0200, Jan Kara wrote:
> On Thu 22-04-10 12:23:29, Miklos Szeredi wrote:
> > I have very little understanding of I/O scheduling but my idea of what's
> > really needed here is to realize that one queue is not able to saturate
> > the device and there's a large backlog of requests on other queues that
> > are waiting to be served. Is something like that implementable?
> I see a problem with defining "saturate the device" - but maybe we could
> measure something like "completed requests / sec" and try autotuning
> slice_idle to maximize this value (hopefully the utility function should
> be concave so we can just use "local optimization").
Yeah, detecting saturation may be difficult.
I guess that function depends on a lot of other things as well,
including seek times, etc. Not easy to optimize.
I'm still wondering what makes such a difference between CFQ on 2.6.16
and CFQ on 2.6.27-34, why is the one in older kernels performing so much
better in this situation?
What should we tell our customers? The default settings should at least
handle these systems a bit better.
Thanks,
Miklos
On Thu, 2010-04-22 at 16:31 -0400, Vivek Goyal wrote:
> On Thu, Apr 22, 2010 at 09:59:14AM +0200, Corrado Zoccolo wrote:
> > Hi Miklos,
> > On Wed, Apr 21, 2010 at 6:05 PM, Miklos Szeredi <[email protected]> wrote:
> > > Jens, Corrado,
> > >
> > > Here's a graph showing the number of issued but not yet completed
> > > requests versus time for CFQ and NOOP schedulers running the tiobench
> > > benchmark with 8 threads:
> > >
> > > http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/queue-depth.jpg
> > >
> > > It shows pretty clearly the performance problem is because CFQ is not
> > > issuing enough request to fill the bandwidth.
> > >
> > > Is this the correct behavior of CFQ or is this a bug?
> > This is the expected behavior from CFQ, even if it is not optimal,
> > since we aren't able to identify multi-splindle disks yet.
>
> In the past we were of the opinion that for sequential workload multi spindle
> disks will not matter much as readahead logic (in OS and possibly in
> hardware also) will help. For random workload we anyway don't idle on the
> single cfqq so it is fine. But my tests now seem to be telling a different
> story.
>
> I also have one FC link to one of the HP EVA and I am running increasing
> number of sequential readers to see if throughput goes up as number of
> readers go up. The results are with noop and cfq. I do flush OS caches
> across the runs but I have no control on caching on HP EVA.
>
> Kernel=2.6.34-rc5
> DIR=/mnt/iostestmnt/fio DEV=/dev/mapper/mpathe
> Workload=bsr iosched=cfq Filesz=2G bs=4K
> =========================================================================
> job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
> --- --- -- ------------ ----------- ------------- -----------
> bsr 1 1 135366 59024 0 0
> bsr 1 2 124256 126808 0 0
> bsr 1 4 132921 341436 0 0
> bsr 1 8 129807 392904 0 0
> bsr 1 16 129988 773991 0 0
>
> Kernel=2.6.34-rc5
> DIR=/mnt/iostestmnt/fio DEV=/dev/mapper/mpathe
> Workload=bsr iosched=noop Filesz=2G bs=4K
> =========================================================================
> job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
> --- --- -- ------------ ----------- ------------- -----------
> bsr 1 1 126187 95272 0 0
> bsr 1 2 185154 72908 0 0
> bsr 1 4 224622 88037 0 0
> bsr 1 8 285416 115592 0 0
> bsr 1 16 348564 156846 0 0
>
These numbers are very similar to what I got.
> So in case of NOOP, throughput shotup to 348MB/s but CFQ reamains more or
> less constat, about 130MB/s.
>
> So atleast in this case, a single sequential CFQ queue is not keeing the
> disk busy enough.
>
> I am wondering why my testing results were different in the past. May be
> it was a different piece of hardware and behavior various across hardware?
Probably. I haven't seen this type of behavior on other hardware.
> Anyway, if that's the case, then we probably need to allow IO from
> multiple sequential readers and keep a watch on throughput. If throughput
> drops then reduce the number of parallel sequential readers. Not sure how
> much of code that is but with multiple cfqq going in parallel, ioprio
> logic will more or less stop working in CFQ (on multi-spindle hardware).
Have you tested on older kernels? Around 2.6.16 it seemed to allow more
parallel reads, but that might have been just accidental (due to I/O
being submitted in a different pattern).
Thanks,
Miklos
On Fri, Apr 23, 2010 at 12:57 PM, Miklos Szeredi <[email protected]> wrote:
> On Thu, 2010-04-22 at 16:31 -0400, Vivek Goyal wrote:
>> On Thu, Apr 22, 2010 at 09:59:14AM +0200, Corrado Zoccolo wrote:
>> > Hi Miklos,
>> > On Wed, Apr 21, 2010 at 6:05 PM, Miklos Szeredi <[email protected]> wrote:
>> > > Jens, Corrado,
>> > >
>> > > Here's a graph showing the number of issued but not yet completed
>> > > requests versus time for CFQ and NOOP schedulers running the tiobench
>> > > benchmark with 8 threads:
>> > >
>> > > http://www.kernel.org/pub/linux/kernel/people/mszeredi/blktrace/queue-depth.jpg
>> > >
>> > > It shows pretty clearly the performance problem is because CFQ is not
>> > > issuing enough request to fill the bandwidth.
>> > >
>> > > Is this the correct behavior of CFQ or is this a bug?
>> > This is the expected behavior from CFQ, even if it is not optimal,
>> > since we aren't able to identify multi-splindle disks yet.
>>
>> In the past we were of the opinion that for sequential workload multi spindle
>> disks will not matter much as readahead logic (in OS and possibly in
>> hardware also) will help. For random workload we anyway don't idle on the
>> single cfqq so it is fine. But my tests now seem to be telling a different
>> story.
>>
>> I also have one FC link to one of the HP EVA and I am running increasing
>> number of sequential readers to see if throughput goes up as number of
>> readers go up. The results are with noop and cfq. I do flush OS caches
>> across the runs but I have no control on caching on HP EVA.
>>
>> Kernel=2.6.34-rc5
>> DIR=/mnt/iostestmnt/fio DEV=/dev/mapper/mpathe
>> Workload=bsr iosched=cfq Filesz=2G bs=4K
>> =========================================================================
>> job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
>> --- --- -- ------------ ----------- ------------- -----------
>> bsr 1 1 135366 59024 0 0
>> bsr 1 2 124256 126808 0 0
>> bsr 1 4 132921 341436 0 0
>> bsr 1 8 129807 392904 0 0
>> bsr 1 16 129988 773991 0 0
>>
>> Kernel=2.6.34-rc5
>> DIR=/mnt/iostestmnt/fio DEV=/dev/mapper/mpathe
>> Workload=bsr iosched=noop Filesz=2G bs=4K
>> =========================================================================
>> job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
>> --- --- -- ------------ ----------- ------------- -----------
>> bsr 1 1 126187 95272 0 0
>> bsr 1 2 185154 72908 0 0
>> bsr 1 4 224622 88037 0 0
>> bsr 1 8 285416 115592 0 0
>> bsr 1 16 348564 156846 0 0
>>
>
> These numbers are very similar to what I got.
>
>> So in case of NOOP, throughput shotup to 348MB/s but CFQ reamains more or
>> less constat, about 130MB/s.
>>
>> So atleast in this case, a single sequential CFQ queue is not keeing the
>> disk busy enough.
>>
>> I am wondering why my testing results were different in the past. May be
>> it was a different piece of hardware and behavior various across hardware?
>
> Probably. I haven't seen this type of behavior on other hardware.
>
>> Anyway, if that's the case, then we probably need to allow IO from
>> multiple sequential readers and keep a watch on throughput. If throughput
>> drops then reduce the number of parallel sequential readers. Not sure how
>> much of code that is but with multiple cfqq going in parallel, ioprio
>> logic will more or less stop working in CFQ (on multi-spindle hardware).
Hi Vivek,
I tried to implement exactly what you are proposing, see the attached patches.
I leverage the queue merging features to let multiple cfqqs share the
disk in the same timeslice.
I changed the queue split code to trigger on throughput drop instead
of on seeky pattern, so diverging queues can remain merged if they
have good throughput. Moreover, I measure the max bandwidth reached by
single queues and merged queues (you can see the values in the
bandwidth sysfs file).
If merged queues can outperform non-merged ones, the queue merging
code will try to opportunistically merge together queues that cannot
submit enough requests to fill half of the NCQ slots. I'd like to know
if you can see any improvements out of this on your hardware. There
are some magic numbers in the code, you may want to try tuning them.
Note that, since the opportunistic queue merging will start happening
only after merged queues have shown to reach higher bandwidth than
non-merged queues, you should use the disk for a while before trying
the test (and you can check sysfs), or the merging will not happen.
>
> Have you tested on older kernels? Around 2.6.16 it seemed to allow more
> parallel reads, but that might have been just accidental (due to I/O
> being submitted in a different pattern).
Is the BW for 1 single reader also better on 2.6.16, or the
improvement is only seen with more concurrent readers?
Thanks,
Corrado
>
> Thanks,
> Miklos
>
>
On Sat, Apr 24, 2010 at 10:36:48PM +0200, Corrado Zoccolo wrote:
[..]
> >
> >> Anyway, if that's the case, then we probably need to allow IO from
> >> multiple sequential readers and keep a watch on throughput. If throughput
> >> drops then reduce the number of parallel sequential readers. Not sure how
> >> much of code that is but with multiple cfqq going in parallel, ioprio
> >> logic will more or less stop working in CFQ (on multi-spindle hardware).
> Hi Vivek,
> I tried to implement exactly what you are proposing, see the attached patches.
> I leverage the queue merging features to let multiple cfqqs share the
> disk in the same timeslice.
> I changed the queue split code to trigger on throughput drop instead
> of on seeky pattern, so diverging queues can remain merged if they
> have good throughput. Moreover, I measure the max bandwidth reached by
> single queues and merged queues (you can see the values in the
> bandwidth sysfs file).
> If merged queues can outperform non-merged ones, the queue merging
> code will try to opportunistically merge together queues that cannot
> submit enough requests to fill half of the NCQ slots. I'd like to know
> if you can see any improvements out of this on your hardware. There
> are some magic numbers in the code, you may want to try tuning them.
> Note that, since the opportunistic queue merging will start happening
> only after merged queues have shown to reach higher bandwidth than
> non-merged queues, you should use the disk for a while before trying
> the test (and you can check sysfs), or the merging will not happen.
Thanks corrado. Using split queue sounds like the right place to do it.
I will test it and report back my results.
>
> >
> > Have you tested on older kernels? ?Around 2.6.16 it seemed to allow more
> > parallel reads, but that might have been just accidental (due to I/O
> > being submitted in a different pattern).
> Is the BW for 1 single reader also better on 2.6.16, or the
> improvement is only seen with more concurrent readers?
I will also test 2.6.16. I am anyway curious, how come 2.6.16 performed
better and we were dispatching requests from multiple cfqq and driving
deeper queue depths. To me this is fundamental cfq design that at one
time one queue gets to use the disk (at least for sync-idle tree). So
something must have been different in 2.6.16.
Thanks
Vivek
On Sat, Apr 24, 2010 at 10:36:48PM +0200, Corrado Zoccolo wrote:
[..]
> >> Anyway, if that's the case, then we probably need to allow IO from
> >> multiple sequential readers and keep a watch on throughput. If throughput
> >> drops then reduce the number of parallel sequential readers. Not sure how
> >> much of code that is but with multiple cfqq going in parallel, ioprio
> >> logic will more or less stop working in CFQ (on multi-spindle hardware).
> Hi Vivek,
> I tried to implement exactly what you are proposing, see the attached patches.
> I leverage the queue merging features to let multiple cfqqs share the
> disk in the same timeslice.
> I changed the queue split code to trigger on throughput drop instead
> of on seeky pattern, so diverging queues can remain merged if they
> have good throughput. Moreover, I measure the max bandwidth reached by
> single queues and merged queues (you can see the values in the
> bandwidth sysfs file).
> If merged queues can outperform non-merged ones, the queue merging
> code will try to opportunistically merge together queues that cannot
> submit enough requests to fill half of the NCQ slots. I'd like to know
> if you can see any improvements out of this on your hardware. There
> are some magic numbers in the code, you may want to try tuning them.
> Note that, since the opportunistic queue merging will start happening
> only after merged queues have shown to reach higher bandwidth than
> non-merged queues, you should use the disk for a while before trying
> the test (and you can check sysfs), or the merging will not happen.
Hi Corrado,
I ran these patches and I did not see any improvement. I think the reason
being that no cooperative queue merging took place and we did not have
any data for throughput with coop flag on.
#cat /sys/block/dm-3/queue/iosched/bandwidth
230 753 0
I think we need to implement something similiar to hw_tag detection logic
where we allow dispatches from multiple sync-idle queues at a time and try
to observe the BW. After certain window once we have observed the window,
then set the system behavior accordingly.
Kernel=2.6.34-rc5-corrado-multicfq
DIR= /mnt/iostmnt/fio DEV= /dev/mapper/mpathe
Workload=bsr iosched=cfq Filesz=2G bs=4K
==========================================================================
job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
--- --- -- ------------ ----------- ------------- -----------
bsr 1 1 126590 61448 0 0
bsr 1 2 127849 242843 0 0
bsr 1 4 131886 508021 0 0
bsr 1 8 131890 398241 0 0
bsr 1 16 129167 454244 0 0
Thanks
Vivek
On Mon, Apr 26, 2010 at 9:14 PM, Vivek Goyal <[email protected]> wrote:
> On Sat, Apr 24, 2010 at 10:36:48PM +0200, Corrado Zoccolo wrote:
>
> [..]
>> >> Anyway, if that's the case, then we probably need to allow IO from
>> >> multiple sequential readers and keep a watch on throughput. If throughput
>> >> drops then reduce the number of parallel sequential readers. Not sure how
>> >> much of code that is but with multiple cfqq going in parallel, ioprio
>> >> logic will more or less stop working in CFQ (on multi-spindle hardware).
>> Hi Vivek,
>> I tried to implement exactly what you are proposing, see the attached patches.
>> I leverage the queue merging features to let multiple cfqqs share the
>> disk in the same timeslice.
>> I changed the queue split code to trigger on throughput drop instead
>> of on seeky pattern, so diverging queues can remain merged if they
>> have good throughput. Moreover, I measure the max bandwidth reached by
>> single queues and merged queues (you can see the values in the
>> bandwidth sysfs file).
>> If merged queues can outperform non-merged ones, the queue merging
>> code will try to opportunistically merge together queues that cannot
>> submit enough requests to fill half of the NCQ slots. I'd like to know
>> if you can see any improvements out of this on your hardware. There
>> are some magic numbers in the code, you may want to try tuning them.
>> Note that, since the opportunistic queue merging will start happening
>> only after merged queues have shown to reach higher bandwidth than
>> non-merged queues, you should use the disk for a while before trying
>> the test (and you can check sysfs), or the merging will not happen.
>
> Hi Corrado,
>
> I ran these patches and I did not see any improvement. I think the reason
> being that no cooperative queue merging took place and we did not have
> any data for throughput with coop flag on.
>
> #cat /sys/block/dm-3/queue/iosched/bandwidth
> 230 753 0
>
> I think we need to implement something similiar to hw_tag detection logic
> where we allow dispatches from multiple sync-idle queues at a time and try
> to observe the BW. After certain window once we have observed the window,
> then set the system behavior accordingly.
Hi Vivek,
thanks for testing. Can you try changing the condition to enable the
queue merging to also consider the case in which max_bw[1] == 0 &&
max_bw[0] > 100MB/s (notice that max_bw is measured in
sectors/jiffie).
This should rule out low end disks, and enable merging where it can be
beneficial.
If the results are good, but we find this enabling condition
unreliable, then we can think of a better way, but I'm curious to see
if the results are promising before thinking to the details.
Thanks,
Corrado
>
> Kernel=2.6.34-rc5-corrado-multicfq
> DIR= /mnt/iostmnt/fio DEV= /dev/mapper/mpathe
> Workload=bsr iosched=cfq Filesz=2G bs=4K
> ==========================================================================
> job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
> --- --- -- ------------ ----------- ------------- -----------
> bsr 1 1 126590 61448 0 0
> bsr 1 2 127849 242843 0 0
> bsr 1 4 131886 508021 0 0
> bsr 1 8 131890 398241 0 0
> bsr 1 16 129167 454244 0 0
>
> Thanks
> Vivek
>
--
__________________________________________________________________________
dott. Corrado Zoccolo mailto:[email protected]
PhD - Department of Computer Science - University of Pisa, Italy
--------------------------------------------------------------------------
The self-confidence of a warrior is not the self-confidence of the average
man. The average man seeks certainty in the eyes of the onlooker and calls
that self-confidence. The warrior seeks impeccability in his own eyes and
calls that humbleness.
Tales of Power - C. Castaneda
On Tue, Apr 27, 2010 at 07:25:14PM +0200, Corrado Zoccolo wrote:
> On Mon, Apr 26, 2010 at 9:14 PM, Vivek Goyal <[email protected]> wrote:
> > On Sat, Apr 24, 2010 at 10:36:48PM +0200, Corrado Zoccolo wrote:
> >
> > [..]
> >> >> Anyway, if that's the case, then we probably need to allow IO from
> >> >> multiple sequential readers and keep a watch on throughput. If throughput
> >> >> drops then reduce the number of parallel sequential readers. Not sure how
> >> >> much of code that is but with multiple cfqq going in parallel, ioprio
> >> >> logic will more or less stop working in CFQ (on multi-spindle hardware).
> >> Hi Vivek,
> >> I tried to implement exactly what you are proposing, see the attached patches.
> >> I leverage the queue merging features to let multiple cfqqs share the
> >> disk in the same timeslice.
> >> I changed the queue split code to trigger on throughput drop instead
> >> of on seeky pattern, so diverging queues can remain merged if they
> >> have good throughput. Moreover, I measure the max bandwidth reached by
> >> single queues and merged queues (you can see the values in the
> >> bandwidth sysfs file).
> >> If merged queues can outperform non-merged ones, the queue merging
> >> code will try to opportunistically merge together queues that cannot
> >> submit enough requests to fill half of the NCQ slots. I'd like to know
> >> if you can see any improvements out of this on your hardware. There
> >> are some magic numbers in the code, you may want to try tuning them.
> >> Note that, since the opportunistic queue merging will start happening
> >> only after merged queues have shown to reach higher bandwidth than
> >> non-merged queues, you should use the disk for a while before trying
> >> the test (and you can check sysfs), or the merging will not happen.
> >
> > Hi Corrado,
> >
> > I ran these patches and I did not see any improvement. I think the reason
> > being that no cooperative queue merging took place and we did not have
> > any data for throughput with coop flag on.
> >
> > #cat /sys/block/dm-3/queue/iosched/bandwidth
> > 230 ? ? 753 ? ? 0
> >
> > I think we need to implement something similiar to hw_tag detection logic
> > where we allow dispatches from multiple sync-idle queues at a time and try
> > to observe the BW. After certain window once we have observed the window,
> > then set the system behavior accordingly.
> Hi Vivek,
> thanks for testing. Can you try changing the condition to enable the
> queue merging to also consider the case in which max_bw[1] == 0 &&
> max_bw[0] > 100MB/s (notice that max_bw is measured in
> sectors/jiffie).
> This should rule out low end disks, and enable merging where it can be
> beneficial.
> If the results are good, but we find this enabling condition
> unreliable, then we can think of a better way, but I'm curious to see
> if the results are promising before thinking to the details.
Ok, I made some changes and now some queue merging seems to be happening
and I am getting little better BW. This will require more debugging. I
will try to spare some time later.
Kernel=2.6.34-rc5-corrado-multicfq
DIR= /mnt/iostmnt/fio DEV= /dev/mapper/mpathe
Workload=bsr iosched=cfq Filesz=1G bs=16K
==========================================================================
job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
--- --- -- ------------ ----------- ------------- -----------
bsr 1 1 136457 67353 0 0
bsr 1 2 148008 192415 0 0
bsr 1 4 180223 535205 0 0
bsr 1 8 166983 542326 0 0
bsr 1 16 176617 832188 0 0
Output of iosched/bandwidth
0 546 740
I did following changes on top of your patch.
Vivek
---
block/cfq-iosched.c | 11 +++++++++--
1 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 4e9e015..7589c44 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -243,6 +243,7 @@ struct cfq_data {
*/
int hw_tag_est_depth;
unsigned int hw_tag_samples;
+ unsigned int cfqq_merged_samples;
/*
* performance measurements
* max_bw is indexed by coop flag.
@@ -1736,10 +1737,14 @@ static struct cfq_queue *cfqq_close(struct cfq_data *cfqd,
// Opportunistic queue merging could be beneficial even on far queues
// We enable it only on NCQ disks, if we observed that merged queues
// can reach higher bandwidth than single queues.
+ // 204 sectors per jiffy is equivalent to 100MB/s on 1000 HZ conf.
+ // Allow merge if we don't have sufficient merged cfqq samples.
rs = cur_cfqq->allocated[READ] + cur_cfqq->allocated[WRITE];
- if (cfqd->hw_tag && cfqd->max_bw[1] > cfqd->max_bw[0] &&
+ if (cfqd->hw_tag
+ && (cfqd->max_bw[1] > cfqd->max_bw[0]
+ || (cfqd->max_bw[0] > 204 && !sample_valid(cfqd->cfqq_merged_samples)))
// Do not overload the device queue
- rs < cfqd->hw_tag_est_depth / 2) {
+ && rs < cfqd->hw_tag_est_depth / 2) {
unsigned r1 = __cfqq->allocated[READ] + __cfqq->allocated[WRITE];
unsigned r2 = __cfqq1->allocated[READ] + __cfqq1->allocated[WRITE];
// Prefer merging with a queue that has fewer pending requests
@@ -1750,6 +1755,8 @@ static struct cfq_queue *cfqq_close(struct cfq_data *cfqd,
// Do not merge if the merged queue would have too many requests
if (r1 + rs > cfqd->hw_tag_est_depth / 2)
return NULL;
+ cfqd->cfqq_merged_samples++;
+
// Merge only if the BW of the two queues is comparable
if (abs(__cfqq->last_bw - cur_cfqq->last_bw) * 4 < cfqd->max_bw[0])
return __cfqq;
--
1.6.2.5
>
> Thanks,
> Corrado
>
> >
> > Kernel=2.6.34-rc5-corrado-multicfq
> > DIR= /mnt/iostmnt/fio ? ? ? ? ?DEV= /dev/mapper/mpathe
> > Workload=bsr ? ? ? iosched=cfq ? ? ?Filesz=2G ? ?bs=4K
> > ==========================================================================
> > job ? ? ? Set NR ?ReadBW(KB/s) ? MaxClat(us) ? ?WriteBW(KB/s) ?MaxClat(us)
> > --- ? ? ? --- -- ?------------ ? ----------- ? ?------------- ?-----------
> > bsr ? ? ? 1 ? 1 ? 126590 ? ? ? ? 61448 ? ? ? ? ?0 ? ? ? ? ? ? ?0
> > bsr ? ? ? 1 ? 2 ? 127849 ? ? ? ? 242843 ? ? ? ? 0 ? ? ? ? ? ? ?0
> > bsr ? ? ? 1 ? 4 ? 131886 ? ? ? ? 508021 ? ? ? ? 0 ? ? ? ? ? ? ?0
> > bsr ? ? ? 1 ? 8 ? 131890 ? ? ? ? 398241 ? ? ? ? 0 ? ? ? ? ? ? ?0
> > bsr ? ? ? 1 ? 16 ?129167 ? ? ? ? 454244 ? ? ? ? 0 ? ? ? ? ? ? ?0
> >
> > Thanks
> > Vivek
> >
>
>
>
> --
> __________________________________________________________________________
>
> dott. Corrado Zoccolo mailto:[email protected]
> PhD - Department of Computer Science - University of Pisa, Italy
> --------------------------------------------------------------------------
> The self-confidence of a warrior is not the self-confidence of the average
> man. The average man seeks certainty in the eyes of the onlooker and calls
> that self-confidence. The warrior seeks impeccability in his own eyes and
> calls that humbleness.
> Tales of Power - C. Castaneda
On Sat, 2010-05-01 at 14:13 +0200, Corrado Zoccolo wrote:
> On Wed, Apr 28, 2010 at 10:02 PM, Vivek Goyal <[email protected]> wrote:
> > On Tue, Apr 27, 2010 at 07:25:14PM +0200, Corrado Zoccolo wrote:
> >> On Mon, Apr 26, 2010 at 9:14 PM, Vivek Goyal <[email protected]> wrote:
> >> > On Sat, Apr 24, 2010 at 10:36:48PM +0200, Corrado Zoccolo wrote:
> >> >
> >> > [..]
> >> >> >> Anyway, if that's the case, then we probably need to allow IO from
> >> >> >> multiple sequential readers and keep a watch on throughput. If throughput
> >> >> >> drops then reduce the number of parallel sequential readers. Not sure how
> >> >> >> much of code that is but with multiple cfqq going in parallel, ioprio
> >> >> >> logic will more or less stop working in CFQ (on multi-spindle hardware).
> >> >> Hi Vivek,
> >> >> I tried to implement exactly what you are proposing, see the attached patches.
> >> >> I leverage the queue merging features to let multiple cfqqs share the
> >> >> disk in the same timeslice.
> >> >> I changed the queue split code to trigger on throughput drop instead
> >> >> of on seeky pattern, so diverging queues can remain merged if they
> >> >> have good throughput. Moreover, I measure the max bandwidth reached by
> >> >> single queues and merged queues (you can see the values in the
> >> >> bandwidth sysfs file).
> >> >> If merged queues can outperform non-merged ones, the queue merging
> >> >> code will try to opportunistically merge together queues that cannot
> >> >> submit enough requests to fill half of the NCQ slots. I'd like to know
> >> >> if you can see any improvements out of this on your hardware. There
> >> >> are some magic numbers in the code, you may want to try tuning them.
> >> >> Note that, since the opportunistic queue merging will start happening
> >> >> only after merged queues have shown to reach higher bandwidth than
> >> >> non-merged queues, you should use the disk for a while before trying
> >> >> the test (and you can check sysfs), or the merging will not happen.
> >> >
> >> > Hi Corrado,
> >> >
> >> > I ran these patches and I did not see any improvement. I think the reason
> >> > being that no cooperative queue merging took place and we did not have
> >> > any data for throughput with coop flag on.
> >> >
> >> > #cat /sys/block/dm-3/queue/iosched/bandwidth
> >> > 230 753 0
> >> >
> >> > I think we need to implement something similiar to hw_tag detection logic
> >> > where we allow dispatches from multiple sync-idle queues at a time and try
> >> > to observe the BW. After certain window once we have observed the window,
> >> > then set the system behavior accordingly.
> >> Hi Vivek,
> >> thanks for testing. Can you try changing the condition to enable the
> >> queue merging to also consider the case in which max_bw[1] == 0 &&
> >> max_bw[0] > 100MB/s (notice that max_bw is measured in
> >> sectors/jiffie).
> >> This should rule out low end disks, and enable merging where it can be
> >> beneficial.
> >> If the results are good, but we find this enabling condition
> >> unreliable, then we can think of a better way, but I'm curious to see
> >> if the results are promising before thinking to the details.
> >
> > Ok, I made some changes and now some queue merging seems to be happening
> > and I am getting little better BW. This will require more debugging. I
> > will try to spare some time later.
> >
> > Kernel=2.6.34-rc5-corrado-multicfq
> > DIR= /mnt/iostmnt/fio DEV= /dev/mapper/mpathe
> > Workload=bsr iosched=cfq Filesz=1G bs=16K
> > ==========================================================================
> > job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
> > --- --- -- ------------ ----------- ------------- -----------
> > bsr 1 1 136457 67353 0 0
> > bsr 1 2 148008 192415 0 0
> > bsr 1 4 180223 535205 0 0
> > bsr 1 8 166983 542326 0 0
> > bsr 1 16 176617 832188 0 0
> >
> > Output of iosched/bandwidth
> >
> > 0 546 740
> >
> > I did following changes on top of your patch.
> This is becoming interesting. I think a major limitation of the
> current approach is that it is too easy for a merged queue to be
> separated again.
> My code:
> if (cfq_cfqq_coop(cfqq) && bw <= cfqd->max_bw[1] * 9/10)
> cfq_mark_cfqq_split_coop(cfqq);
> will immediately split any merged queue as soon as max_bw[1] grows too
> much, so it should be based on max_bw[0].
> Moreover, this code will likely split off all cics from the merged
> queue, while it would be much better to split off only the cics that
> are receiving less than their fair share of the BW (this will also
> improve the fairness of the scheduler when queues are merged).
Is there any update on the status of this issue?
Thanks,
Miklos
> Corrado
> >
> > Vivek
> >
> > ---
> > block/cfq-iosched.c | 11 +++++++++--
> > 1 files changed, 9 insertions(+), 2 deletions(-)
> >
> > diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
> > index 4e9e015..7589c44 100644
> > --- a/block/cfq-iosched.c
> > +++ b/block/cfq-iosched.c
> > @@ -243,6 +243,7 @@ struct cfq_data {
> > */
> > int hw_tag_est_depth;
> > unsigned int hw_tag_samples;
> > + unsigned int cfqq_merged_samples;
> > /*
> > * performance measurements
> > * max_bw is indexed by coop flag.
> > @@ -1736,10 +1737,14 @@ static struct cfq_queue *cfqq_close(struct cfq_data *cfqd,
> > // Opportunistic queue merging could be beneficial even on far queues
> > // We enable it only on NCQ disks, if we observed that merged queues
> > // can reach higher bandwidth than single queues.
> > + // 204 sectors per jiffy is equivalent to 100MB/s on 1000 HZ conf.
> > + // Allow merge if we don't have sufficient merged cfqq samples.
> > rs = cur_cfqq->allocated[READ] + cur_cfqq->allocated[WRITE];
> > - if (cfqd->hw_tag && cfqd->max_bw[1] > cfqd->max_bw[0] &&
> > + if (cfqd->hw_tag
> > + && (cfqd->max_bw[1] > cfqd->max_bw[0]
> > + || (cfqd->max_bw[0] > 204 && !sample_valid(cfqd->cfqq_merged_samples)))
> > // Do not overload the device queue
> > - rs < cfqd->hw_tag_est_depth / 2) {
> > + && rs < cfqd->hw_tag_est_depth / 2) {
> > unsigned r1 = __cfqq->allocated[READ] + __cfqq->allocated[WRITE];
> > unsigned r2 = __cfqq1->allocated[READ] + __cfqq1->allocated[WRITE];
> > // Prefer merging with a queue that has fewer pending requests
> > @@ -1750,6 +1755,8 @@ static struct cfq_queue *cfqq_close(struct cfq_data *cfqd,
> > // Do not merge if the merged queue would have too many requests
> > if (r1 + rs > cfqd->hw_tag_est_depth / 2)
> > return NULL;
> > + cfqd->cfqq_merged_samples++;
> > +
> > // Merge only if the BW of the two queues is comparable
> > if (abs(__cfqq->last_bw - cur_cfqq->last_bw) * 4 < cfqd->max_bw[0])
> > return __cfqq;
> > --
> > 1.6.2.5
> >
> >>
> >> Thanks,
> >> Corrado
> >>
> >> >
> >> > Kernel=2.6.34-rc5-corrado-multicfq
> >> > DIR= /mnt/iostmnt/fio DEV= /dev/mapper/mpathe
> >> > Workload=bsr iosched=cfq Filesz=2G bs=4K
> >> > ==========================================================================
> >> > job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
> >> > --- --- -- ------------ ----------- ------------- -----------
> >> > bsr 1 1 126590 61448 0 0
> >> > bsr 1 2 127849 242843 0 0
> >> > bsr 1 4 131886 508021 0 0
> >> > bsr 1 8 131890 398241 0 0
> >> > bsr 1 16 129167 454244 0 0
> >> >
> >> > Thanks
> >> > Vivek
> >> >
> >>
> >>
> >>
> >> --
> >> __________________________________________________________________________
> >>
> >> dott. Corrado Zoccolo mailto:[email protected]
> >> PhD - Department of Computer Science - University of Pisa, Italy
> >> --------------------------------------------------------------------------
> >> The self-confidence of a warrior is not the self-confidence of the average
> >> man. The average man seeks certainty in the eyes of the onlooker and calls
> >> that self-confidence. The warrior seeks impeccability in his own eyes and
> >> calls that humbleness.
> >> Tales of Power - C. Castaneda
> >
>
>
>
On Mon, Jun 14, 2010 at 07:59:13PM +0200, Miklos Szeredi wrote:
> On Sat, 2010-05-01 at 14:13 +0200, Corrado Zoccolo wrote:
> > On Wed, Apr 28, 2010 at 10:02 PM, Vivek Goyal <[email protected]> wrote:
> > > On Tue, Apr 27, 2010 at 07:25:14PM +0200, Corrado Zoccolo wrote:
> > >> On Mon, Apr 26, 2010 at 9:14 PM, Vivek Goyal <[email protected]> wrote:
> > >> > On Sat, Apr 24, 2010 at 10:36:48PM +0200, Corrado Zoccolo wrote:
> > >> >
> > >> > [..]
> > >> >> >> Anyway, if that's the case, then we probably need to allow IO from
> > >> >> >> multiple sequential readers and keep a watch on throughput. If throughput
> > >> >> >> drops then reduce the number of parallel sequential readers. Not sure how
> > >> >> >> much of code that is but with multiple cfqq going in parallel, ioprio
> > >> >> >> logic will more or less stop working in CFQ (on multi-spindle hardware).
> > >> >> Hi Vivek,
> > >> >> I tried to implement exactly what you are proposing, see the attached patches.
> > >> >> I leverage the queue merging features to let multiple cfqqs share the
> > >> >> disk in the same timeslice.
> > >> >> I changed the queue split code to trigger on throughput drop instead
> > >> >> of on seeky pattern, so diverging queues can remain merged if they
> > >> >> have good throughput. Moreover, I measure the max bandwidth reached by
> > >> >> single queues and merged queues (you can see the values in the
> > >> >> bandwidth sysfs file).
> > >> >> If merged queues can outperform non-merged ones, the queue merging
> > >> >> code will try to opportunistically merge together queues that cannot
> > >> >> submit enough requests to fill half of the NCQ slots. I'd like to know
> > >> >> if you can see any improvements out of this on your hardware. There
> > >> >> are some magic numbers in the code, you may want to try tuning them.
> > >> >> Note that, since the opportunistic queue merging will start happening
> > >> >> only after merged queues have shown to reach higher bandwidth than
> > >> >> non-merged queues, you should use the disk for a while before trying
> > >> >> the test (and you can check sysfs), or the merging will not happen.
> > >> >
> > >> > Hi Corrado,
> > >> >
> > >> > I ran these patches and I did not see any improvement. I think the reason
> > >> > being that no cooperative queue merging took place and we did not have
> > >> > any data for throughput with coop flag on.
> > >> >
> > >> > #cat /sys/block/dm-3/queue/iosched/bandwidth
> > >> > 230 753 0
> > >> >
> > >> > I think we need to implement something similiar to hw_tag detection logic
> > >> > where we allow dispatches from multiple sync-idle queues at a time and try
> > >> > to observe the BW. After certain window once we have observed the window,
> > >> > then set the system behavior accordingly.
> > >> Hi Vivek,
> > >> thanks for testing. Can you try changing the condition to enable the
> > >> queue merging to also consider the case in which max_bw[1] == 0 &&
> > >> max_bw[0] > 100MB/s (notice that max_bw is measured in
> > >> sectors/jiffie).
> > >> This should rule out low end disks, and enable merging where it can be
> > >> beneficial.
> > >> If the results are good, but we find this enabling condition
> > >> unreliable, then we can think of a better way, but I'm curious to see
> > >> if the results are promising before thinking to the details.
> > >
> > > Ok, I made some changes and now some queue merging seems to be happening
> > > and I am getting little better BW. This will require more debugging. I
> > > will try to spare some time later.
> > >
> > > Kernel=2.6.34-rc5-corrado-multicfq
> > > DIR= /mnt/iostmnt/fio DEV= /dev/mapper/mpathe
> > > Workload=bsr iosched=cfq Filesz=1G bs=16K
> > > ==========================================================================
> > > job Set NR ReadBW(KB/s) MaxClat(us) WriteBW(KB/s) MaxClat(us)
> > > --- --- -- ------------ ----------- ------------- -----------
> > > bsr 1 1 136457 67353 0 0
> > > bsr 1 2 148008 192415 0 0
> > > bsr 1 4 180223 535205 0 0
> > > bsr 1 8 166983 542326 0 0
> > > bsr 1 16 176617 832188 0 0
> > >
> > > Output of iosched/bandwidth
> > >
> > > 0 546 740
> > >
> > > I did following changes on top of your patch.
> > This is becoming interesting. I think a major limitation of the
> > current approach is that it is too easy for a merged queue to be
> > separated again.
> > My code:
> > if (cfq_cfqq_coop(cfqq) && bw <= cfqd->max_bw[1] * 9/10)
> > cfq_mark_cfqq_split_coop(cfqq);
> > will immediately split any merged queue as soon as max_bw[1] grows too
> > much, so it should be based on max_bw[0].
> > Moreover, this code will likely split off all cics from the merged
> > queue, while it would be much better to split off only the cics that
> > are receiving less than their fair share of the BW (this will also
> > improve the fairness of the scheduler when queues are merged).
>
> Is there any update on the status of this issue?
How about running cfq with slice_idle=0 on high end storage. This should make
it very close to deadline behavior?
There has not been any further progress on my end for merging more
sequential queues for achieving better throughput.
Thanks
Vivek
