On 06/29/2015 05:21 PM, Bruce Cran wrote:
On Linux, using the latest fio sources, "fio
examples/tiobench-example.fio" results in the progress indicator
(percentage and ETA) displaying odd values. Percentage rises to about
70% before dropping to 20%, getting to 25% and then the test finishes.
The ETA starts out about 4 minutes, falls to 1m and then jumps to 12.
$ ./fio examples/tiobench-example.fio
[snip]
f4: Laying out IO file(s) (1 file(s) / 512MB)
Jobs: 4 (f=4): [_(12),r(4)] [25.1% done] [887KB/0KB/0KB /s] [221/0/0
iops] [eta 12m:00s]
f1: (groupid=0, jobs=1): err= 0: pid=9730: Mon Jun 29 17:17:01 2015
write: io=388340KB, bw=6453.7KB/s, iops=1613, runt= 60174msec
clat (usec): min=144, max=704183, avg=616.69, stdev=7110.74
lat (usec): min=145, max=704184, avg=617.10, stdev=7110.74
clat percentiles (usec):
| 1.00th=[ 354], 5.00th=[ 378], 10.00th=[ 394], 20.00th=[ 410],
| 30.00th=[ 422], 40.00th=[ 434], 50.00th=[ 442], 60.00th=[ 450],
| 70.00th=[ 466], 80.00th=[ 478], 90.00th=[ 506], 95.00th=[ 564],
| 99.00th=[ 1416], 99.50th=[ 1448], 99.90th=[24960], 99.95th=[56576],
| 99.99th=[544768]
bw (KB /s): min= 2067, max= 9432, per=26.50%, avg=6839.64,
stdev=1755.81
lat (usec) : 250=0.01%, 500=88.36%, 750=7.53%, 1000=0.02%
lat (msec) : 2=3.91%, 4=0.03%, 10=0.01%, 20=0.01%, 50=0.06%
lat (msec) : 100=0.03%, 250=0.01%, 750=0.01%
I'll take a look at this. Is this a regression?
--
Jens Axboe
--
To unsubscribe from this list: send the line "unsubscribe fio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
