stress-ng classes not being transformed into stressors in job mode

Bug #1706970 reported by Colin Ian King
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Stress-ng
Fix Released
Medium
Colin Ian King

Bug Description

I'm just catching up, on the latest release of stress-ng (0.08.09) (BTW: is there a mailing list I can subscribe to so I get notified of new versions?) and like the look of the --job option to supply a jobfile. I'm obviously missing something fundamental though; the manual says it essentially a file containing stress-ng options without the leading -- but I can't get that to translate.

FWIW I'm running on RHEL7.3.

Here's a sample run from the command line I use fairly often:

$ ./stress-ng --class vm --sequential -1 --timeout 10s --log-brief --metrics-brief
dispatching hogs: 32 bigheap, 32 brk, 32 madvise, 32 malloc, 32 mlock, 32 mmap, 32 mmapfork, 32 mmapmany, 32 mremap, 32 msync, 32 shm, 32 shm-sysv, 32 stack, 32 stackmmap, 32 tmpfs, 32 userfaultfd, 32 vm, 32 vm-rw, 32 vm-splice
successful run completed in 201.35s (3 mins, 21.35 secs)
stressor bogo ops real time usr time sys time bogo ops/s bogo ops/s
                          (secs) (secs) (secs) (real time) (usr+sys time)
bigheap 2048143 12.06 9.24 135.60 169813.48 14140.73
brk 42621714 12.18 7.01 153.85 3499663.15 264961.54
madvise 3682 10.07 2.68 265.58 365.82 13.73
malloc 1349005562 10.00 318.31 0.41 134834348.52 4232572.67
mlock 560106 10.11 0.26 24.16 55393.07 22936.36
mmap 5393 10.00 109.54 75.04 539.28 29.22
mmapfork 107 10.29 93.65 246.82 10.40 0.31
mmapmany 62774392 10.01 7.74 312.04 6271514.06 196304.93
mremap 51917 10.00 0.34 318.95 5191.10 162.60
msync 1468318 10.15 11.02 30.98 144691.96 34959.95
shm 155555 10.01 12.15 306.26 15536.93 488.54
shm-sysv 33997 10.00 11.55 238.84 3398.34 135.78
stack 0 12.85 9.36 129.46 0.00 0.00
stackmmap 56746 10.00 1.88 124.39 5674.48 449.40
tmpfs 939 10.03 183.47 89.94 93.57 3.43
userfaultfd 40365174 10.00 6.45 136.78 4036137.78 281820.67
vm 1633184 10.00 318.34 1.14 163306.45 5112.01
vm-rw 1936670 10.00 0.40 318.73 193661.28 6068.59
vm-splice 67172379 10.00 2.83 316.74 6717202.16 210196.14
$

If I pick up those options and put into a jobfile:

$ cat jobfile1
#
# Jobfile 1 for stress-ng
#
run sequential
class vm
sequential -1
timeout 10s
log-brief
metrics-brief
$

$ ./stress-ng --job jobfile1
No stress workers
$

I'm not sure of the relationship between the 'run sequential' command and the '--sequential -1' command-line option (I didn't get a different result when I tried 'run sequential -1') but my addled brain is obviously misfiring somewhere.

Any advice appreciated...

-Andy

Revision history for this message
Colin Ian King (colin-king) wrote :

Thanks for reporting this. This is indeed a bug and I will get around to fixing it when I get back from my vacation.

Changed in stress-ng:
importance: Undecided → Medium
assignee: nobody → Colin Ian King (colin-king)
status: New → Confirmed
Revision history for this message
Colin Ian King (colin-king) wrote :

I've pushed some fixes to help resolve this issue and it should be landing in ubuntu Artful in the next 24 hours.

With the changes, the following jobfile will do the trick:

#
# Jobfile 1 for stress-ng
#
timeout 2s
log-brief
metrics-brief

sequential 1
verbose
class vm

Changed in stress-ng:
status: Confirmed → Fix Committed
Changed in stress-ng:
status: Fix Committed → Fix Released
Revision history for this message
Andy Williams (wilandy) wrote :

Thanks Colin,

Initial testing looks good. I'll be able to test further over the next couple of days.

-Andy

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Related questions

Remote bug watches

Bug watches keep track of this bug in other bug trackers.