Thanks. On a separate (or may be related) note, shared files handling
in dup_files seems to be broken (unless I'm missing the logic behind
dropping non-shared files). The patch below has fixed it for me.
diff --git a/init.c b/init.c
index 9b0666d..99038b2 100644
--- a/init.c
+++ b/init.c
@@ -1160,17 +1161,20 @@ static int add_job(struct thread_data *td, const char *j
td_new->o.new_group = 0;
if (file_alloced) {
- td_new->files_index = 0;
- td_new->files_size = 0;
if (td_new->files) {
struct fio_file *f;
for_each_file(td_new, f, i) {
if (f->file_name)
- free(f->file_name);
- free(f);
+ sfree(f->file_name);
+ sfree(f);
}
+ free(td_new->files);
td_new->files = NULL;
}
+ td_new->files_index = 0;
+ td_new->files_size = 0;
if (td_new->o.filename) {
free(td_new->o.filename);
td_new->o.filename = NULL;
Regards,
Andrey
On Tue, Apr 1, 2014 at 7:31 AM, Jens Axboe <[email protected]> wrote:
> On 03/31/2014 07:08 PM, Andrey Kuzmin wrote:
>>
>> The following simple setup
>> ---
>> [global]
>> ioengine=psync
>> direct=1
>> sync=1
>>
>> [randrwrite]
>> filename=/tmp/test.dat
>> lockfile=exclusive
>> filesize=1m
>> ;
>> thread=1
>> numjobs=4
>> runtime=10
>> time_based=1
>> ;
>> rw=randwrite
>> bs=4k
>> ---
>>
>> runs fine with jobs being processes, and fails as follows with thread=1.
>>
>> randrwrite: (g=0): rw=randwrite, bs=4K-4K/4K-4K/4K-4K, ioengine=psync,
>> iodepth=1
>> ...
>> fio-2.1.6.1-23-g9f97
>> Starting 4 threads
>> fio: pid=10455, err=9/file:engines/sync.c:67, func=xfer, error=Bad
>> file descriptor]
>> fio: pid=10453, err=9/file:engines/sync.c:67, func=xfer, error=Bad
>> file descriptor
>> fio: pid=10454, err=9/file:engines/sync.c:67, func=xfer, error=Bad
>> file descriptor
>> Jobs: 1 (f=1): [wXXX] [100.0% done] [0KB/2372KB/0KB /s] [0/593/0 iops]
>> [eta 00m:00s]
>> randrwrite: (groupid=0, jobs=1): err= 0: pid=10452: Mon Mar 31 18:05:26
>> 2014
>> write: io=14900KB, bw=1489.9KB/s, iops=372, runt= 10001msec
>> clat (usec): min=419, max=260739, avg=1546.00, stdev=4329.54
>> lat (usec): min=424, max=260744, avg=1551.95, stdev=4329.58
>> clat percentiles (usec):
>> | 1.00th=[ 588], 5.00th=[ 732], 10.00th=[ 780], 20.00th=[
>> 828],
>> | 30.00th=[ 876], 40.00th=[ 956], 50.00th=[ 1080], 60.00th=[
>> 1432],
>> | 70.00th=[ 2064], 80.00th=[ 2256], 90.00th=[ 2480], 95.00th=[
>> 2736],
>> | 99.00th=[ 3344], 99.50th=[ 3632], 99.90th=[ 4512],
>> 99.95th=[23680],
>> | 99.99th=[261120]
>> bw (KB /s): min= 5, max= 1099, per=22.09%, avg=519.90,
>> stdev=359.34
>> lat (usec) : 500=0.19%, 750=6.42%, 1000=37.61%
>> lat (msec) : 2=23.68%, 4=31.81%, 10=0.24%, 50=0.03%, 500=0.03%
>> cpu : usr=0.76%, sys=8.08%, ctx=8191, majf=0, minf=5
>> 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=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=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,
>> >=64=0.0%
>> issued : total=r=0/w=3725/d=0, short=r=0/w=0/d=0
>> latency : target=0, window=0, percentile=100.00%, depth=1
>> randrwrite: (groupid=0, jobs=1): err= 9 (file:engines/sync.c:67,
>> func=xfer, error=Bad file descriptor): pid=10453: Mon Mar 31 18:05:26
>> 2014
>> write: io=3072.0KB, bw=581465B/s, iops=142, runt= 5410msec
>> clat (usec): min=587, max=33827, avg=1817.50, stdev=2464.14
>> lat (usec): min=593, max=33837, avg=1823.85, stdev=2464.47
>> clat percentiles (usec):
>> | 1.00th=[ 804], 5.00th=[ 876], 10.00th=[ 932], 20.00th=[
>> 988],
>> | 30.00th=[ 1048], 40.00th=[ 1096], 50.00th=[ 1208], 60.00th=[
>> 2024],
>> | 70.00th=[ 2224], 80.00th=[ 2352], 90.00th=[ 2576], 95.00th=[
>> 2896],
>> | 99.00th=[ 4080], 99.50th=[32640], 99.90th=[34048],
>> 99.95th=[34048],
>> | 99.99th=[34048]
>> bw (KB /s): min= 165, max= 1266, per=25.28%, avg=595.00,
>> stdev=450.04
>> lat (usec) : 750=0.39%, 1000=22.76%
>> lat (msec) : 2=36.67%, 4=38.88%, 10=0.52%, 20=0.13%, 50=0.52%
>> cpu : usr=0.15%, sys=5.69%, ctx=2288, majf=0, minf=0
>> 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=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,
>> >=64=0.0%
>> complete : 0=0.1%, 4=99.9%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,
>> >=64=0.0%
>> issued : total=r=0/w=769/d=0, short=r=0/w=0/d=0
>> latency : target=0, window=0, percentile=100.00%, depth=1
>> randrwrite: (groupid=0, jobs=1): err= 9 (file:engines/sync.c:67,
>> func=xfer, error=Bad file descriptor): pid=10454: Mon Mar 31 18:05:26
>> 2014
>> write: io=2528.0KB, bw=478497B/s, iops=117, runt= 5410msec
>> clat (usec): min=299, max=33920, avg=2106.79, stdev=3775.38
>> lat (usec): min=303, max=33930, avg=2113.60, stdev=3775.68
>> clat percentiles (usec):
>> | 1.00th=[ 564], 5.00th=[ 820], 10.00th=[ 932], 20.00th=[
>> 996],
>> | 30.00th=[ 1064], 40.00th=[ 1160], 50.00th=[ 1256], 60.00th=[
>> 1976],
>> | 70.00th=[ 2224], 80.00th=[ 2384], 90.00th=[ 2672], 95.00th=[
>> 3056],
>> | 99.00th=[33024], 99.50th=[33024], 99.90th=[34048],
>> 99.95th=[34048],
>> | 99.99th=[34048]
>> bw (KB /s): min= 46, max= 728, per=18.61%, avg=438.17,
>> stdev=265.37
>> lat (usec) : 500=0.16%, 750=3.63%, 1000=16.59%
>> lat (msec) : 2=39.81%, 4=37.28%, 10=0.63%, 20=0.32%, 50=1.42%
>> cpu : usr=0.30%, sys=5.03%, ctx=2048, majf=0, minf=2
>> 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=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,
>> >=64=0.0%
>> complete : 0=0.2%, 4=99.8%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,
>> >=64=0.0%
>> issued : total=r=0/w=633/d=0, short=r=0/w=0/d=0
>> latency : target=0, window=0, percentile=100.00%, depth=1
>> randrwrite: (groupid=0, jobs=1): err= 9 (file:engines/sync.c:67,
>> func=xfer, error=Bad file descriptor): pid=10455: Mon Mar 31 18:05:26
>> 2014
>> write: io=3052.0KB, bw=583286B/s, iops=142, runt= 5358msec
>> clat (usec): min=541, max=24712, avg=1707.90, stdev=1533.18
>> lat (usec): min=547, max=24743, avg=1714.89, stdev=1533.64
>> clat percentiles (usec):
>> | 1.00th=[ 796], 5.00th=[ 876], 10.00th=[ 908], 20.00th=[
>> 980],
>> | 30.00th=[ 1048], 40.00th=[ 1128], 50.00th=[ 1240], 60.00th=[
>> 1736],
>> | 70.00th=[ 2224], 80.00th=[ 2384], 90.00th=[ 2640], 95.00th=[
>> 2896],
>> | 99.00th=[ 3280], 99.50th=[ 4640], 99.90th=[24704],
>> 99.95th=[24704],
>> | 99.99th=[24704]
>> bw (KB /s): min= 204, max= 1009, per=20.39%, avg=480.00,
>> stdev=283.93
>> lat (usec) : 750=0.65%, 1000=22.12%
>> lat (msec) : 2=38.35%, 4=38.09%, 10=0.26%, 50=0.39%
>> cpu : usr=0.37%, sys=5.45%, ctx=2261, majf=0, minf=5
>> 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=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,
>> >=64=0.0%
>> complete : 0=0.1%, 4=99.9%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%,
>> >=64=0.0%
>> issued : total=r=0/w=764/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):
>> WRITE: io=23552KB, aggrb=2354KB/s, minb=467KB/s, maxb=1489KB/s,
>> mint=5358msec, maxt=10001msec
>>
>> Disk stats (read/write):
>> sda: ios=0/13818, merge=0/2227, ticks=0/7428, in_queue=7260,
>> util=72.37%
>>
>> Is file sharing with non-trivial lockfile= setting is expected to work
>> with threads=1, or is it an unanticipated use case?
>
>
> No, that's supposed to work. I think it's a race between shared files, and
> opening/closing of them when you have a timed based job. I'll take a look at
> it.
>
> --
> Jens Axboe
>
--
To unsubscribe from this list: send the line "unsubscribe fio" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
