ceph version : master (ceph version 0.80-713-g86754cc (867= 54cc78ca570f19f5a68fb634d613f952a22eb))

fio version :=C2=A0fio-2.1.9-20= -g290a

gdb backtrace

#0 =C2=A00x000= 07ffff6de5249 in AO_fetch_and_add_full (incr=3D1, p=3D0x7fff00000018) at /u= sr/include/atomic_ops/sysdeps/gcc/x86.h:68

#1 =C2=A0inc (this=3D0x7fff00000018) at ./include/atomic.h:98

#2 =C2=A0ceph::buffer::ptr::ptr (this=3D0x7fffecf74820, p=3D..., o=3D<= optimized out>, l=3D0) at common/buffer.cc:575

#3 =C2=A00x0000= 7ffff6de63df in ceph::buffer::list::append (this=3Dthis@entry=3D0x7fffc8000= 8e8, bp=3D..., off=3D<optimized out>, len=3Dlen@entry=3D0) at common/= buffer.cc:1267

#4 =C2=A00x00007ffff6de6a44 in ceph::buffer::list::splice (this=3D0x7f= ffe4014590, off=3D<optimized out>, len=3D64512, claim_by=3D0x7fffc800= 08e8) at common/buffer.cc:1426

#5 =C2=A00x00007ffff7b89d45 in Str= iper::StripedReadResult::add_partial_sparse_result (this=3D0x7fffe4014278, = cct=3D0x7fffe4006f50, bl=3D..., bl_map=3D..., bl_off=3D3670016,

=C2=A0 =C2=A0 buffer_extents=3D...) at osdc/Striper.cc:291

#= 6 =C2=A00x00007ffff7b180d8 in librbd::C_AioRead::finish (this=3D0x7fffe400d= 6a0, r=3D<optimized out>) at librbd/AioCompletion.cc:94

#7 = =C2=A00x00007ffff7b182f9 in Context::complete (this=3D0x7fffe400d6a0, r=3D&= lt;optimized out>) at ./include/Context.h:64

#8 =C2=A00x00007ffff7b1840d in librbd::AioRequest::complete (this=3D0x= 7fffe4014540, r=3D0) at ./librbd/AioRequest.h:40

#9 =C2=A00x00007= ffff6d3a538 in librados::C_AioComplete::finish (this=3D0x7fffdc0025c0, r=3D= <optimized out>) at ./librados/AioCompletionImpl.h:178

#10 0x00007ffff7b182f9 in Context::complete (this=3D0x7fffdc0025c0, r= =3D<optimized out>) at ./include/Context.h:64

#11 0x00007ff= ff6dc85f0 in Finisher::finisher_thread_entry (this=3D0x7fffe400c7c8) at com= mon/Finisher.cc:56

#12 0x00007ffff5f49f8e in start_thread (arg=3D0x7fffecf75700) at pthre= ad_create.c:311

#13 0x00007ffff5a6fa0d in clone () at ../sysdeps/= unix/sysv/linux/x86_64/clone.S:113

On Tue, Jun 3, 2014 at 12:34 PM, Danny Al-Gaaf = <danny.al-g= aaf-2YacvwyR+KOzQB+pC5nmwQ@public.gmane.org> wrote:

Am 03.06.2014 20:55, schrieb Sushma R:

> Haomai,
>
> I'm using the latest ceph master branch.
>
> ceph_smalliobench is a Ceph internal benchmarking tool similar to rado= s
> bench and the performance is more or less similar to that reported by = fio.
>
> I tried to use fio with rbd ioengine (
> http://telekomcloud.github.io/= ceph/2014/02/26/ceph-performance-analysis_fio_rbd.html)
> and below are the numbers with different workloads on our setup.
> Note : fio rbd engine segfaults with randread IO pattern, only with Le= velDB
> (no issues with FileStore). With FileStore, performance of
> ceph_smalliobench and fio-rbd is similar for READs, so the numbers for=
> randread for LevelDB are with ceph_smalliobench (since fio rbd segfaul= ts).

Could you send me a backtrace of the segfault and some info about the=
ceph and fio version you used so that I can take a look at it?

Thanks,

Danny