* fio terse output for WRITE not working.
@ 2014-08-08 18:49 Steve Wise
2014-08-15 14:35 ` Steve Wise
0 siblings, 1 reply; 8+ messages in thread
From: Steve Wise @ 2014-08-08 18:49 UTC (permalink / raw)
To: fio
Hey,
I'm using fio-2.1.10, and trying to get the Total IO via the terse output. It works for
read but not for write. Here is the output and the fio conf file I'm using.
Any thoughts? Is this a known bug?
Thanks,
Steve.
-----
Read works:
[root@hpc1 ~]# RW=randread fio --minimal fio-test.conf
3;fio-2.1.10;foo;0;0;80251;40105;40105;2001;4;85;8.170228;4.004539;65;805;189.872799;49.78
1808;1.000000%=122;5.000000%=133;10.000000%=139;20.000000%=149;30.000000%=159;40.000000%=1
69;50.000000%=179;60.000000%=191;70.000000%=205;80.000000%=225;90.000000%=258;95.000000%=2
86;99.000000%=354;99.500000%=386;99.900000%=450;99.950000%=486;99.990000%=596;0%=0;0%=0;0%
=0;87;809;198.185493;49.501590;38468;42140;100.000000%;40262.666667;1837.395258;0;0;0;0;0;
0;0.000000;0.000000;0;0;0.000000;0.000000;1.000000%=0;5.000000%=0;10.000000%=0;20.000000%=
0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.000000%=0;80.000000%=0;90.000000%
=0;95.000000%=0;99.000000%=0;99.500000%=0;99.900000%=0;99.950000%=0;99.990000%=0;0%=0;0%=0
;0%=0;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0.000000;7.750000%;41.150000%;31556;0;9
;0.1%;0.1%;0.1%;100.0%;0.0%;0.0%;0.0%;0.00%;0.00%;0.00%;0.00%;0.00%;0.01%;88.31%;11.64%;0.
04%;0.01%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%
Write shows zeros for Total IO:
[root@hpc1 ~]# RW=randwrite fio --minimal fio-test.conf
3;fio-2.1.10;foo;0;0;0;0;0;0;0;0;0.000000;0.000000;0;0;0.000000;0.000000;1.000000%=0;5.000
000%=0;10.000000%=0;20.000000%=0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.00
0000%=0;80.000000%=0;90.000000%=0;95.000000%=0;99.000000%=0;99.500000%=0;99.900000%=0;99.9
50000%=0;99.990000%=0;0%=0;0%=0;0%=0;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0.000000
;5409;2697;2697;2005;6;51;18.330190;7.389300;81;10033;2939.422629;1572.743671;1.000000%=88
4;5.000000%=1144;10.000000%=1320;20.000000%=1560;30.000000%=1784;40.000000%=2096;50.000000
%=2448;60.000000%=3120;70.000000%=3664;80.000000%=4320;90.000000%=5216;95.000000%=5984;99.
000000%=7520;99.500000%=8032;99.900000%=9664;99.950000%=9792;99.990000%=10048;0%=0;0%=0;0%
=0;89;10040;2957.946016;1571.794210;2572;2794;100.000000%;2701.000000;95.951377;2.894212%;
8.582834%;4987;0;8;0.1%;0.1%;0.1%;99.9%;0.0%;0.0%;0.0%;0.00%;0.00%;0.00%;0.00%;0.00%;0.17%
;0.17%;0.02%;0.07%;2.02%;34.72%;36.75%;26.07%;0.02%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.
00%;0.00%
Without --minimal, we see Total IO WRITE numbers:
[root@hpc1 ~]# RW=randwrite fio fio-test.conf
foo: (g=0): rw=randwrite, bs=1K-1K/1K-1K/1K-1K, ioengine=libaio, iodepth=8
fio-2.1.10
Starting 1 process
Jobs: 1 (f=1): [w] [-.-% done] [0KB/1158KB/0KB /s] [0/1158/0 iops] [eta 00m:00s]
foo: (groupid=0, jobs=1): err= 0: pid=4547: Fri Aug 8 11:45:00 2014
write: io=2334.0KB, bw=1162.1KB/s, iops=1162, runt= 2007msec
slat (usec): min=7, max=45, avg=19.02, stdev= 4.98
clat (usec): min=72, max=15679, avg=6849.86, stdev=3321.20
lat (usec): min=81, max=15699, avg=6869.15, stdev=3322.08
clat percentiles (usec):
| 1.00th=[ 1864], 5.00th=[ 2960], 10.00th=[ 3408], 20.00th=[ 3824],
| 30.00th=[ 4192], 40.00th=[ 4704], 50.00th=[ 5728], 60.00th=[ 7904],
| 70.00th=[ 9152], 80.00th=[10176], 90.00th=[11712], 95.00th=[12864],
| 99.00th=[13888], 99.50th=[14400], 99.90th=[15424], 99.95th=[15424],
| 99.99th=[15680]
bw (KB /s): min= 1101, max= 1226, per=99.91%, avg=1161.00, stdev=60.88
lat (usec) : 100=0.30%, 250=0.43%
lat (msec) : 2=0.39%, 4=21.34%, 10=55.70%, 20=21.85%
cpu : usr=1.60%, sys=4.34%, ctx=2304, majf=0, minf=8
IO depths : 1=0.1%, 2=0.1%, 4=0.2%, 8=99.7%, 16=0.0%, 32=0.0%, >=64=0.0%
submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
complete : 0=0.0%, 4=100.0%, 8=0.1%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
issued : total=r=0/w=2334/d=0, short=r=0/w=0/d=0
latency : target=0, window=0, percentile=100.00%, depth=8
Run status group 0 (all jobs):
WRITE: io=2334KB, aggrb=1162KB/s, minb=1162KB/s, maxb=1162KB/s, mint=2007msec,
maxt=2007msec
Disk stats (read/write):
sda: ios=0/2207, merge=0/0, ticks=0/15041, in_queue=15076, util=95.05%
Here is the conf file:
[root@hpc1 ~]# cat fio-test.conf
[foo]
rw=${RW}
size=40m
bs=1k
directory=/tmp
direct=1
numjobs=1
runtime=2
time_based
ioengine=libaio
iodepth=8
^ permalink raw reply [flat|nested] 8+ messages in thread* RE: fio terse output for WRITE not working. 2014-08-08 18:49 fio terse output for WRITE not working Steve Wise @ 2014-08-15 14:35 ` Steve Wise 2014-08-15 15:28 ` Vasily Tarasov 0 siblings, 1 reply; 8+ messages in thread From: Steve Wise @ 2014-08-15 14:35 UTC (permalink / raw) To: fio Any thoughts on this? Thanks, Steve. > -----Original Message----- > From: fio-owner@vger.kernel.org [mailto:fio-owner@vger.kernel.org] On Behalf Of Steve > Wise > Sent: Friday, August 08, 2014 1:49 PM > To: fio@vger.kernel.org > Subject: fio terse output for WRITE not working. > > Hey, > > I'm using fio-2.1.10, and trying to get the Total IO via the terse output. It works for > read but not for write. Here is the output and the fio conf file I'm using. > > Any thoughts? Is this a known bug? > > Thanks, > > Steve. > > ----- > > > Read works: > > [root@hpc1 ~]# RW=randread fio --minimal fio-test.conf > 3;fio-2.1.10;foo;0;0;80251;40105;40105;2001;4;85;8.170228;4.004539;65;805;189.872799;49.78 > 1808;1.000000%=122;5.000000%=133;10.000000%=139;20.000000%=149;30.000000%=159;40. > 000000%=1 > 69;50.000000%=179;60.000000%=191;70.000000%=205;80.000000%=225;90.000000%=258;95. > 000000%=2 > 86;99.000000%=354;99.500000%=386;99.900000%=450;99.950000%=486;99.990000%=596;0% > =0;0%=0;0% > =0;87;809;198.185493;49.501590;38468;42140;100.000000%;40262.666667;1837.395258;0;0;0; > 0;0; > 0;0.000000;0.000000;0;0;0.000000;0.000000;1.000000%=0;5.000000%=0;10.000000%=0;20.000 > 000%= > 0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.000000%=0;80.000000%=0;9 > 0.000000% > =0;95.000000%=0;99.000000%=0;99.500000%=0;99.900000%=0;99.950000%=0;99.990000%=0; > 0%=0;0%=0 > ;0%=0;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0.000000;7.750000%;41.150000%;31556 > ;0;9 > ;0.1%;0.1%;0.1%;100.0%;0.0%;0.0%;0.0%;0.00%;0.00%;0.00%;0.00%;0.00%;0.01%;88.31%;11. > 64%;0. > 04%;0.01%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00% > > Write shows zeros for Total IO: > > [root@hpc1 ~]# RW=randwrite fio --minimal fio-test.conf > 3;fio-2.1.10;foo;0;0;0;0;0;0;0;0;0.000000;0.000000;0;0;0.000000;0.000000;1.000000%=0;5.000 > 000%=0;10.000000%=0;20.000000%=0;30.000000%=0;40.000000%=0;50.000000%=0;60.00000 > 0%=0;70.00 > 0000%=0;80.000000%=0;90.000000%=0;95.000000%=0;99.000000%=0;99.500000%=0;99.9000 > 00%=0;99.9 > 50000%=0;99.990000%=0;0%=0;0%=0;0%=0;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0 > .000000 > ;5409;2697;2697;2005;6;51;18.330190;7.389300;81;10033;2939.422629;1572.743671;1.000000% > =88 > 4;5.000000%=1144;10.000000%=1320;20.000000%=1560;30.000000%=1784;40.000000%=2096; > 50.000000 > %=2448;60.000000%=3120;70.000000%=3664;80.000000%=4320;90.000000%=5216;95.000000 > %=5984;99. > 000000%=7520;99.500000%=8032;99.900000%=9664;99.950000%=9792;99.990000%=10048;0 > %=0;0%=0;0% > =0;89;10040;2957.946016;1571.794210;2572;2794;100.000000%;2701.000000;95.951377;2.8942 > 12%; > 8.582834%;4987;0;8;0.1%;0.1%;0.1%;99.9%;0.0%;0.0%;0.0%;0.00%;0.00%;0.00%;0.00%;0.00% > ;0.17% > ;0.17%;0.02%;0.07%;2.02%;34.72%;36.75%;26.07%;0.02%;0.00%;0.00%;0.00%;0.00%;0.00%;0. > 00%;0. > 00%;0.00% > > Without --minimal, we see Total IO WRITE numbers: > > [root@hpc1 ~]# RW=randwrite fio fio-test.conf > foo: (g=0): rw=randwrite, bs=1K-1K/1K-1K/1K-1K, ioengine=libaio, iodepth=8 > fio-2.1.10 > Starting 1 process > Jobs: 1 (f=1): [w] [-.-% done] [0KB/1158KB/0KB /s] [0/1158/0 iops] [eta 00m:00s] > foo: (groupid=0, jobs=1): err= 0: pid=4547: Fri Aug 8 11:45:00 2014 > write: io=2334.0KB, bw=1162.1KB/s, iops=1162, runt= 2007msec > slat (usec): min=7, max=45, avg=19.02, stdev= 4.98 > clat (usec): min=72, max=15679, avg=6849.86, stdev=3321.20 > lat (usec): min=81, max=15699, avg=6869.15, stdev=3322.08 > clat percentiles (usec): > | 1.00th=[ 1864], 5.00th=[ 2960], 10.00th=[ 3408], 20.00th=[ 3824], > | 30.00th=[ 4192], 40.00th=[ 4704], 50.00th=[ 5728], 60.00th=[ 7904], > | 70.00th=[ 9152], 80.00th=[10176], 90.00th=[11712], 95.00th=[12864], > | 99.00th=[13888], 99.50th=[14400], 99.90th=[15424], 99.95th=[15424], > | 99.99th=[15680] > bw (KB /s): min= 1101, max= 1226, per=99.91%, avg=1161.00, stdev=60.88 > lat (usec) : 100=0.30%, 250=0.43% > lat (msec) : 2=0.39%, 4=21.34%, 10=55.70%, 20=21.85% > cpu : usr=1.60%, sys=4.34%, ctx=2304, majf=0, minf=8 > IO depths : 1=0.1%, 2=0.1%, 4=0.2%, 8=99.7%, 16=0.0%, 32=0.0%, >=64=0.0% > submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% > complete : 0=0.0%, 4=100.0%, 8=0.1%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% > issued : total=r=0/w=2334/d=0, short=r=0/w=0/d=0 > latency : target=0, window=0, percentile=100.00%, depth=8 > > Run status group 0 (all jobs): > WRITE: io=2334KB, aggrb=1162KB/s, minb=1162KB/s, maxb=1162KB/s, mint=2007msec, > maxt=2007msec > > Disk stats (read/write): > sda: ios=0/2207, merge=0/0, ticks=0/15041, in_queue=15076, util=95.05% > > > Here is the conf file: > > [root@hpc1 ~]# cat fio-test.conf > [foo] > rw=${RW} > size=40m > bs=1k > directory=/tmp > direct=1 > numjobs=1 > runtime=2 > time_based > ioengine=libaio > iodepth=8 > > -- > To unsubscribe from this list: send the line "unsubscribe fio" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: fio terse output for WRITE not working. 2014-08-15 14:35 ` Steve Wise @ 2014-08-15 15:28 ` Vasily Tarasov 2014-08-15 15:31 ` Steve Wise 0 siblings, 1 reply; 8+ messages in thread From: Vasily Tarasov @ 2014-08-15 15:28 UTC (permalink / raw) To: Steve Wise; +Cc: fio@vger.kernel.org Hi Steve, I found terse/minimal output to be quite unstable and inconsistent in fio. So, as suggested by Jens, I switched to json output (--output-format=json). It seems to be more stable and is also quite easy to parse. Perhaps that's an option for you. Vasily On Fri, Aug 15, 2014 at 10:35 AM, Steve Wise <swise@opengridcomputing.com> wrote: > Any thoughts on this? > > Thanks, > > Steve. > >> -----Original Message----- >> From: fio-owner@vger.kernel.org [mailto:fio-owner@vger.kernel.org] On Behalf Of Steve >> Wise >> Sent: Friday, August 08, 2014 1:49 PM >> To: fio@vger.kernel.org >> Subject: fio terse output for WRITE not working. >> >> Hey, >> >> I'm using fio-2.1.10, and trying to get the Total IO via the terse output. It works for >> read but not for write. Here is the output and the fio conf file I'm using. >> >> Any thoughts? Is this a known bug? >> >> Thanks, >> >> Steve. >> >> ----- >> >> >> Read works: >> >> [root@hpc1 ~]# RW=randread fio --minimal fio-test.conf >> > 3;fio-2.1.10;foo;0;0;80251;40105;40105;2001;4;85;8.170228;4.004539;65;805;189.872799;49.78 >> 1808;1.000000%=122;5.000000%=133;10.000000%=139;20.000000%=149;30.000000%=159;40. >> 000000%=1 >> 69;50.000000%=179;60.000000%=191;70.000000%=205;80.000000%=225;90.000000%=258;95. >> 000000%=2 >> 86;99.000000%=354;99.500000%=386;99.900000%=450;99.950000%=486;99.990000%=596;0% >> =0;0%=0;0% >> =0;87;809;198.185493;49.501590;38468;42140;100.000000%;40262.666667;1837.395258;0;0;0; >> 0;0; >> 0;0.000000;0.000000;0;0;0.000000;0.000000;1.000000%=0;5.000000%=0;10.000000%=0;20.000 >> 000%= >> 0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.000000%=0;80.000000%=0;9 >> 0.000000% >> =0;95.000000%=0;99.000000%=0;99.500000%=0;99.900000%=0;99.950000%=0;99.990000%=0; >> 0%=0;0%=0 >> ;0%=0;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0.000000;7.750000%;41.150000%;31556 >> ;0;9 >> ;0.1%;0.1%;0.1%;100.0%;0.0%;0.0%;0.0%;0.00%;0.00%;0.00%;0.00%;0.00%;0.01%;88.31%;11. >> 64%;0. >> 04%;0.01%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00% >> >> Write shows zeros for Total IO: >> >> [root@hpc1 ~]# RW=randwrite fio --minimal fio-test.conf >> > 3;fio-2.1.10;foo;0;0;0;0;0;0;0;0;0.000000;0.000000;0;0;0.000000;0.000000;1.000000%=0;5.000 >> 000%=0;10.000000%=0;20.000000%=0;30.000000%=0;40.000000%=0;50.000000%=0;60.00000 >> 0%=0;70.00 >> 0000%=0;80.000000%=0;90.000000%=0;95.000000%=0;99.000000%=0;99.500000%=0;99.9000 >> 00%=0;99.9 >> 50000%=0;99.990000%=0;0%=0;0%=0;0%=0;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0 >> .000000 >> ;5409;2697;2697;2005;6;51;18.330190;7.389300;81;10033;2939.422629;1572.743671;1.000000% >> =88 >> 4;5.000000%=1144;10.000000%=1320;20.000000%=1560;30.000000%=1784;40.000000%=2096; >> 50.000000 >> %=2448;60.000000%=3120;70.000000%=3664;80.000000%=4320;90.000000%=5216;95.000000 >> %=5984;99. >> 000000%=7520;99.500000%=8032;99.900000%=9664;99.950000%=9792;99.990000%=10048;0 >> %=0;0%=0;0% >> =0;89;10040;2957.946016;1571.794210;2572;2794;100.000000%;2701.000000;95.951377;2.8942 >> 12%; >> 8.582834%;4987;0;8;0.1%;0.1%;0.1%;99.9%;0.0%;0.0%;0.0%;0.00%;0.00%;0.00%;0.00%;0.00% >> ;0.17% >> ;0.17%;0.02%;0.07%;2.02%;34.72%;36.75%;26.07%;0.02%;0.00%;0.00%;0.00%;0.00%;0.00%;0. >> 00%;0. >> 00%;0.00% >> >> Without --minimal, we see Total IO WRITE numbers: >> >> [root@hpc1 ~]# RW=randwrite fio fio-test.conf >> foo: (g=0): rw=randwrite, bs=1K-1K/1K-1K/1K-1K, ioengine=libaio, iodepth=8 >> fio-2.1.10 >> Starting 1 process >> Jobs: 1 (f=1): [w] [-.-% done] [0KB/1158KB/0KB /s] [0/1158/0 iops] [eta 00m:00s] >> foo: (groupid=0, jobs=1): err= 0: pid=4547: Fri Aug 8 11:45:00 2014 >> write: io=2334.0KB, bw=1162.1KB/s, iops=1162, runt= 2007msec >> slat (usec): min=7, max=45, avg=19.02, stdev= 4.98 >> clat (usec): min=72, max=15679, avg=6849.86, stdev=3321.20 >> lat (usec): min=81, max=15699, avg=6869.15, stdev=3322.08 >> clat percentiles (usec): >> | 1.00th=[ 1864], 5.00th=[ 2960], 10.00th=[ 3408], 20.00th=[ 3824], >> | 30.00th=[ 4192], 40.00th=[ 4704], 50.00th=[ 5728], 60.00th=[ 7904], >> | 70.00th=[ 9152], 80.00th=[10176], 90.00th=[11712], 95.00th=[12864], >> | 99.00th=[13888], 99.50th=[14400], 99.90th=[15424], 99.95th=[15424], >> | 99.99th=[15680] >> bw (KB /s): min= 1101, max= 1226, per=99.91%, avg=1161.00, stdev=60.88 >> lat (usec) : 100=0.30%, 250=0.43% >> lat (msec) : 2=0.39%, 4=21.34%, 10=55.70%, 20=21.85% >> cpu : usr=1.60%, sys=4.34%, ctx=2304, majf=0, minf=8 >> IO depths : 1=0.1%, 2=0.1%, 4=0.2%, 8=99.7%, 16=0.0%, 32=0.0%, >=64=0.0% >> submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% >> complete : 0=0.0%, 4=100.0%, 8=0.1%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% >> issued : total=r=0/w=2334/d=0, short=r=0/w=0/d=0 >> latency : target=0, window=0, percentile=100.00%, depth=8 >> >> Run status group 0 (all jobs): >> WRITE: io=2334KB, aggrb=1162KB/s, minb=1162KB/s, maxb=1162KB/s, mint=2007msec, >> maxt=2007msec >> >> Disk stats (read/write): >> sda: ios=0/2207, merge=0/0, ticks=0/15041, in_queue=15076, util=95.05% >> >> >> Here is the conf file: >> >> [root@hpc1 ~]# cat fio-test.conf >> [foo] >> rw=${RW} >> size=40m >> bs=1k >> directory=/tmp >> direct=1 >> numjobs=1 >> runtime=2 >> time_based >> ioengine=libaio >> iodepth=8 >> >> -- >> To unsubscribe from this list: send the line "unsubscribe fio" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html > > -- > To unsubscribe from this list: send the line "unsubscribe fio" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > ^ permalink raw reply [flat|nested] 8+ messages in thread
* RE: fio terse output for WRITE not working. 2014-08-15 15:28 ` Vasily Tarasov @ 2014-08-15 15:31 ` Steve Wise 2014-08-22 22:27 ` Jens Axboe 0 siblings, 1 reply; 8+ messages in thread From: Steve Wise @ 2014-08-15 15:31 UTC (permalink / raw) To: 'Vasily Tarasov'; +Cc: fio > -----Original Message----- > From: fio-owner@vger.kernel.org [mailto:fio-owner@vger.kernel.org] On Behalf Of Vasily > Tarasov > Sent: Friday, August 15, 2014 10:29 AM > To: Steve Wise > Cc: fio@vger.kernel.org > Subject: Re: fio terse output for WRITE not working. > > Hi Steve, > > I found terse/minimal output to be quite unstable and inconsistent in > fio. So, as suggested by Jens, I switched to json output > (--output-format=json). It seems to be more stable and is also quite > easy to parse. > > Perhaps that's an option for you. > > Vasily I'll check out json. Thanks Vasily. ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: fio terse output for WRITE not working. 2014-08-15 15:31 ` Steve Wise @ 2014-08-22 22:27 ` Jens Axboe 2014-08-22 23:23 ` Steve Wise 0 siblings, 1 reply; 8+ messages in thread From: Jens Axboe @ 2014-08-22 22:27 UTC (permalink / raw) To: Steve Wise, 'Vasily Tarasov'; +Cc: fio On 2014-08-15 10:31, Steve Wise wrote: > > >> -----Original Message----- >> From: fio-owner@vger.kernel.org [mailto:fio-owner@vger.kernel.org] On Behalf Of Vasily >> Tarasov >> Sent: Friday, August 15, 2014 10:29 AM >> To: Steve Wise >> Cc: fio@vger.kernel.org >> Subject: Re: fio terse output for WRITE not working. >> >> Hi Steve, >> >> I found terse/minimal output to be quite unstable and inconsistent in >> fio. So, as suggested by Jens, I switched to json output >> (--output-format=json). It seems to be more stable and is also quite >> easy to parse. >> >> Perhaps that's an option for you. >> >> Vasily > > I'll check out json. Thanks Vasily. json is vastly superior - easy to parse by both humans and computers. That said, the terse format should of course work. I ran a quick test just now, and I don't see anything wrong? -- Jens Axboe ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: fio terse output for WRITE not working. 2014-08-22 22:27 ` Jens Axboe @ 2014-08-22 23:23 ` Steve Wise 2014-08-22 23:34 ` Jens Axboe 0 siblings, 1 reply; 8+ messages in thread From: Steve Wise @ 2014-08-22 23:23 UTC (permalink / raw) To: Jens Axboe, 'Vasily Tarasov'; +Cc: fio On 8/22/2014 5:27 PM, Jens Axboe wrote: > On 2014-08-15 10:31, Steve Wise wrote: >> >> >>> -----Original Message----- >>> From: fio-owner@vger.kernel.org [mailto:fio-owner@vger.kernel.org] >>> On Behalf Of Vasily >>> Tarasov >>> Sent: Friday, August 15, 2014 10:29 AM >>> To: Steve Wise >>> Cc: fio@vger.kernel.org >>> Subject: Re: fio terse output for WRITE not working. >>> >>> Hi Steve, >>> >>> I found terse/minimal output to be quite unstable and inconsistent in >>> fio. So, as suggested by Jens, I switched to json output >>> (--output-format=json). It seems to be more stable and is also quite >>> easy to parse. >>> >>> Perhaps that's an option for you. >>> >>> Vasily >> >> I'll check out json. Thanks Vasily. > > json is vastly superior - easy to parse by both humans and computers. > That said, the terse format should of course work. I ran a quick test > just now, and I don't see anything wrong? > See my original posting on this. The write terse output only has zeros for the over bandwidth in the first few fields of the record. read works fine for me though. ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: fio terse output for WRITE not working. 2014-08-22 23:23 ` Steve Wise @ 2014-08-22 23:34 ` Jens Axboe 2014-08-22 23:46 ` Steve Wise 0 siblings, 1 reply; 8+ messages in thread From: Jens Axboe @ 2014-08-22 23:34 UTC (permalink / raw) To: Steve Wise, 'Vasily Tarasov'; +Cc: fio On 2014-08-22 17:23, Steve Wise wrote: > On 8/22/2014 5:27 PM, Jens Axboe wrote: >> On 2014-08-15 10:31, Steve Wise wrote: >>> >>> >>>> -----Original Message----- >>>> From: fio-owner@vger.kernel.org [mailto:fio-owner@vger.kernel.org] >>>> On Behalf Of Vasily >>>> Tarasov >>>> Sent: Friday, August 15, 2014 10:29 AM >>>> To: Steve Wise >>>> Cc: fio@vger.kernel.org >>>> Subject: Re: fio terse output for WRITE not working. >>>> >>>> Hi Steve, >>>> >>>> I found terse/minimal output to be quite unstable and inconsistent in >>>> fio. So, as suggested by Jens, I switched to json output >>>> (--output-format=json). It seems to be more stable and is also quite >>>> easy to parse. >>>> >>>> Perhaps that's an option for you. >>>> >>>> Vasily >>> >>> I'll check out json. Thanks Vasily. >> >> json is vastly superior - easy to parse by both humans and computers. >> That said, the terse format should of course work. I ran a quick test >> just now, and I don't see anything wrong? >> > > See my original posting on this. The write terse output only has zeros > for the over bandwidth in the first few fields of the record. read works > fine for me though. Yes, I'm saying I don't see this happening. This is my quickie test case: 3;fio-2.1.11-19-g0297;null;0;0;21019592;7006530;1751632;3000;0;0;0.000000;0.000000;0;31;0.028879;0.181874;1.000000%=0;5.000000%=0;10.000000%=0;20.000000%=0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.000000%=0;80.000000%=0;90.000000%=0;95.000000%=0;99.000000%=1;99.500000%=1;99.900000%=1;99.950000%=1;99.990000%=1;0%=0;0%=0;0%=0;0;32;0.055929;0.249507;5932824;7322384;99.011636%;6937280.000000;569503.420436;21014252;7004750;1751187;3000;0;0;0.000000;0.000000;0;48;0.029822;0.186289;1.000000%=0;5.000000%=0;10.000000%=0;20.000000%=0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.000000%=0;80.000000%=0;90.000000%=0;95.000000%=0;99.000000%=1;99.500000%=1;99.900000%=1;99.950000%=1;99.990000%=1;0%=0;0%=0;0%=0;0;48;0.074062;0.286370;5932824;7324072;98.994494%;6934316.800000;568074.804612;99.666667%;0.000000%;4;0;3;100.0%;0.0%;0.0%;0.0%;0.0%;0.0%;0.0%;100.00%;0.01%;0.01%;0.01%;0.01%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%; 0.00% or when run with normal output: null: (groupid=0, jobs=1): err= 0: pid=26423: Fri Aug 22 17:33:46 2014 read : io=20499MB, bw=6833.2MB/s, iops=1749.3K, runt= 3000msec clat (usec): min=0, max=82, avg= 0.03, stdev= 0.19 lat (usec): min=0, max=83, avg= 0.06, stdev= 0.26 clat percentiles (usec): | 1.00th=[ 0], 5.00th=[ 0], 10.00th=[ 0], 20.00th=[ 0], | 30.00th=[ 0], 40.00th=[ 0], 50.00th=[ 0], 60.00th=[ 0], | 70.00th=[ 0], 80.00th=[ 0], 90.00th=[ 0], 95.00th=[ 0], | 99.00th=[ 1], 99.50th=[ 1], 99.90th=[ 1], 99.95th=[ 1], | 99.99th=[ 1] bw (MB /s): min= 5617, max= 7119, per=98.96%, avg=6762.01, stdev=641.65 write: io=20495MB, bw=6831.6MB/s, iops=1748.9K, runt= 3000msec clat (usec): min=0, max=144, avg= 0.03, stdev= 0.20 lat (usec): min=0, max=144, avg= 0.07, stdev= 0.30 clat percentiles (usec): | 1.00th=[ 0], 5.00th=[ 0], 10.00th=[ 0], 20.00th=[ 0], | 30.00th=[ 0], 40.00th=[ 0], 50.00th=[ 0], 60.00th=[ 0], | 70.00th=[ 0], 80.00th=[ 0], 90.00th=[ 0], 95.00th=[ 0], | 99.00th=[ 1], 99.50th=[ 1], 99.90th=[ 1], 99.95th=[ 1], | 99.99th=[ 1] bw (MB /s): min= 5617, max= 7118, per=98.94%, avg=6759.24, stdev=640.28 lat (usec) : 2=100.00%, 4=0.01%, 10=0.01%, 20=0.01%, 50=0.01% lat (usec) : 100=0.01%, 250=0.01% cpu : usr=100.00%, sys=0.00%, ctx=7, majf=0, minf=3 IO depths : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0% submit : 0=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% complete : 0=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% issued : total=r=5247840/w=5246653/d=0, short=r=0/w=0/d=0 latency : target=0, window=0, percentile=100.00%, depth=1 Run status group 0 (all jobs): READ: io=20499MB, aggrb=6833.2MB/s, minb=6833.2MB/s, maxb=6833.2MB/s, mint=3000msec, maxt=3000msec WRITE: io=20495MB, aggrb=6831.6MB/s, minb=6831.6MB/s, maxb=6831.6MB/s, mint=3000msec, maxt=3000msec -- Jens Axboe ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: fio terse output for WRITE not working. 2014-08-22 23:34 ` Jens Axboe @ 2014-08-22 23:46 ` Steve Wise 0 siblings, 0 replies; 8+ messages in thread From: Steve Wise @ 2014-08-22 23:46 UTC (permalink / raw) To: Jens Axboe, 'Vasily Tarasov'; +Cc: fio > > Yes, I'm saying I don't see this happening. This is my quickie test case: > > 3;fio-2.1.11-19-g0297;null;0;0;21019592;7006530;1751632;3000;0;0;0.000000;0.000000;0;31;0.028879;0.181874;1.000000%=0;5.000000%=0;10.000000%=0;20.000000%=0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.000000%=0;80.000000%=0;90.000000%=0;95.000000%=0;99.000000%=1;99.500000%=1;99.900000%=1;99.950000%=1;99.990000%=1;0%=0;0%=0;0%=0;0;32;0.055929;0.249507;5932824;7322384;99.011636%;6937280.000000;569503.420436;21014252;7004750;1751187;3000;0;0;0.000000;0.000000;0;48;0.029822;0.186289;1.000000%=0;5.000000%=0;10.000000%=0;20.000000%=0;30.000000%=0;40.000000%=0;50.000000%=0;60.000000%=0;70.000000%=0;80.000000%=0;90.000000%=0;95.000000%=0;99.000000%=1;99.500000%=1;99.900000%=1;99.950000%=1;99.990000%=1;0%=0;0%=0;0%=0;0;48;0.074062;0.286370;5932824;7324072;98.994494%;6934316.800000;568074.804612;99.666667%;0.000000%;4;0;3;100.0%;0.0%;0.0%;0.0%;0.0%;0.0%;0.0%;100.00%;0.01%;0.01%;0.01%;0.01%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%;0.00%; > > 0.00% > > or when run with normal output: > > null: (groupid=0, jobs=1): err= 0: pid=26423: Fri Aug 22 17:33:46 2014 > read : io=20499MB, bw=6833.2MB/s, iops=1749.3K, runt= 3000msec > clat (usec): min=0, max=82, avg= 0.03, stdev= 0.19 > lat (usec): min=0, max=83, avg= 0.06, stdev= 0.26 > clat percentiles (usec): > | 1.00th=[ 0], 5.00th=[ 0], 10.00th=[ 0], 20.00th=[ > 0], > | 30.00th=[ 0], 40.00th=[ 0], 50.00th=[ 0], 60.00th=[ > 0], > | 70.00th=[ 0], 80.00th=[ 0], 90.00th=[ 0], 95.00th=[ > 0], > | 99.00th=[ 1], 99.50th=[ 1], 99.90th=[ 1], 99.95th=[ > 1], > | 99.99th=[ 1] > bw (MB /s): min= 5617, max= 7119, per=98.96%, avg=6762.01, > stdev=641.65 > write: io=20495MB, bw=6831.6MB/s, iops=1748.9K, runt= 3000msec > clat (usec): min=0, max=144, avg= 0.03, stdev= 0.20 > lat (usec): min=0, max=144, avg= 0.07, stdev= 0.30 > clat percentiles (usec): > | 1.00th=[ 0], 5.00th=[ 0], 10.00th=[ 0], 20.00th=[ > 0], > | 30.00th=[ 0], 40.00th=[ 0], 50.00th=[ 0], 60.00th=[ > 0], > | 70.00th=[ 0], 80.00th=[ 0], 90.00th=[ 0], 95.00th=[ > 0], > | 99.00th=[ 1], 99.50th=[ 1], 99.90th=[ 1], 99.95th=[ > 1], > | 99.99th=[ 1] > bw (MB /s): min= 5617, max= 7118, per=98.94%, avg=6759.24, > stdev=640.28 > lat (usec) : 2=100.00%, 4=0.01%, 10=0.01%, 20=0.01%, 50=0.01% > lat (usec) : 100=0.01%, 250=0.01% > cpu : usr=100.00%, sys=0.00%, ctx=7, majf=0, minf=3 > IO depths : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, > >=64=0.0% > submit : 0=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, > >=64=0.0% > complete : 0=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, > >=64=0.0% > issued : total=r=5247840/w=5246653/d=0, short=r=0/w=0/d=0 > latency : target=0, window=0, percentile=100.00%, depth=1 > > Run status group 0 (all jobs): > READ: io=20499MB, aggrb=6833.2MB/s, minb=6833.2MB/s, > maxb=6833.2MB/s, mint=3000msec, maxt=3000msec > WRITE: io=20495MB, aggrb=6831.6MB/s, minb=6831.6MB/s, > maxb=6831.6MB/s, mint=3000msec, maxt=3000ms hmm. I'll try with the latest fio from the git repo. Perhaps this has been fixed since the version I'm using (2.1.10)? Thanks, Steve. ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2014-08-22 23:46 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2014-08-08 18:49 fio terse output for WRITE not working Steve Wise 2014-08-15 14:35 ` Steve Wise 2014-08-15 15:28 ` Vasily Tarasov 2014-08-15 15:31 ` Steve Wise 2014-08-22 22:27 ` Jens Axboe 2014-08-22 23:23 ` Steve Wise 2014-08-22 23:34 ` Jens Axboe 2014-08-22 23:46 ` Steve Wise
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox