From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from merlin.infradead.org ([205.233.59.134]:57656 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752796Ab2DNOrM (ORCPT ); Sat, 14 Apr 2012 10:47:12 -0400 Message-ID: <4F898DEC.2000807@kernel.dk> Date: Sat, 14 Apr 2012 16:47:08 +0200 From: Jens Axboe MIME-Version: 1.0 Subject: Re: Recent changes References: <20120406040004.1BF99484001@kernel.dk> <4F7E713E.7030300@suse.de> <4F7EF043.9000305@kernel.dk> <4F7F2A75.6030508@suse.de> <4F7F2B97.7000703@kernel.dk> <4F891D18.5020202@suse.de> In-Reply-To: <4F891D18.5020202@suse.de> Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Sender: fio-owner@vger.kernel.org List-Id: fio@vger.kernel.org To: Danny Kukawka Cc: "fio@vger.kernel.org" -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2012-04-14 08:45, Danny Kukawka wrote: > Am 06.04.2012 19:44, schrieb Jens Axboe: >> On 2012-04-06 11:40, Danny Kukawka wrote: >>> Am 06.04.2012 15:31, schrieb Jens Axboe: >>>> On 2012-04-05 22:29, Danny Kukawka wrote: >>>>> Am 06.04.2012 06:00, schrieb Jens Axboe: >>>>>> The following changes since commit 885ac623a4f154007efa49266bb381bcbc60f1e6: >>>>>> >>>>>> iolog: remove assert in io_u overlap (2012-04-04 14:11:58 -0600) >>>>>> >>>>>> are available in the git repository at: >>>>>> git://git.kernel.dk/fio.git master >>>>>> >>>>>> Jens Axboe (1): >>>>>> group reporting: fix bad values of min/max >>>>> >>>>> Should this fix also the minb/maxb and mint/maxt ? If so, it doesn't >>>>> work for me. >>>>> >>>>> fio --name=fio1-write-4m-rbd --direct=1 --size=50g --group_reporting >>>>> --bs=4m --rw=write --iodepth=128 --ioengine=libaio --filename=/dev/rbd0 >>>>> >>>>> still prints: >>>>> >>>>> READ: io=65260MB, aggrb=111334KB/s, minb=114006KB/s, maxb=114006KB/s, >>>>> mint=600232msec, maxt=600232msec >>>> >>>> Since that's still off by 1.024, are you absolutely sure you are running >>>> the right version of fio? >> >>> I used git snapshot with latest commit >>> ddb754dbb54a8174080f753c85ff56a7ffca3957, but I can recheck if needed. >> >> Yes, please re-check, it should not be happening and I don't see it >> here. >> > > I checked again with 8423bd1106a0 and I still see the same problem: > > fio --name=fio1-write-4m-rbd --direct=1 --size=50g --group_reporting > --bs=4m --rw=write --iodepth=128 --ioengine=libaio > --filename=$TEST_TARGET --output=$DIR/fio1-write-4m-rbd > > Run status group 0 (all jobs): > WRITE: io=51200MB, aggrb=20080KB/s, minb=20080KB/s, maxb=20080KB/s, > mint=2610946msec, maxt=2610946msec I'm confused on what you think is an issue? From the above, fio transferred 51200MB, which is 52428800KB. The runtime was 2610.946s, which gives us a throughput of 20080.38KB/sec. Which is what fio reports. aggrb/minb/maxb are identical, which was the original bug. What am I missing? - -- Jens Axboe -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQIcBAEBCAAGBQJPiY3oAAoJEPfTWPspceCm7TIP/jCGLv0L6Th7T4FosTrCC0bm U9jUDWkNo+b4n+87FLTmX6rxXenNN/NbGYJHaCE8A+FnWyDLihhEAslMD9mqn7OA 66EWir/HrbzikQ6mF9fMPDWCLLNKVV+eVORmrN3Ne6y6d3luH5RiMFKOfWmdbhv2 cHhVpDx5hfGF+tvRcPy/NIOkl7gNszOqF2ili9wP+aTokp+2jF7rVK9pw7KfA9jG IPrt2oC6hIr6+6v8xekZRgEH61R/Gsq6T0Tf0dJtEMMNhcspEpXZbKCf1xV8+22W HmYsjVJysIPEn0KCnPzE+DtDEF/rInbYvlTuOcJS1ugMt/AGOWtQm9OprIQTQsc+ T0lMQ4lVEzxmcvy3dMs2pnTL0pdWOAVfNXGopVpd1jV0jI1uBvsZyW2sBpiVn7nA 10GB6Y97Cm+ap9oaP+nnxRgLjmHrDpbgSFnRt6CFUnD+6xp2iV8Bo7qA06e0ABwT IFpeHW57OhhR+zIMqNqsldvsuuPB8PKWxvi0eH7YBuDKk/XoD91Im9iYJRxjbvBQ xN7hpjeWFQap4OLPERKDoTYBcJ4NdYaM9UdXPsmazc4T28K2tmxtzcIrkalj1Ryn d6B0FoQRWZlPYk3c7NbDKcz49llC08zdptNffMP2g8fKzgGxzTY/oAgiET6JqK/o dSMKbZgMjfK2iNGUjLR9 =Mpll -----END PGP SIGNATURE-----