Message-ID: <20170103221510.dkrxqav7qrq772ji@thunk.org>
Date:   Tue, 3 Jan 2017 17:15:10 -0500
From:   Theodore Ts'o <tytso@....edu>
To:     Roman Penyaev <roman.penyaev@...fitbricks.com>
Cc:     Namjae Jeon <namjae.jeon@...sung.com>,
        Andreas Dilger <adilger.kernel@...ger.ca>,
        linux-ext4@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/3] ext4: Find desired extent in
 ext4_ext_shift_extents() using binsearch

On Tue, Jan 03, 2017 at 09:44:15PM +0100, Roman Penyaev wrote:
> 
> (I had to say that right now I am testing on 4.4.28 kernel and testing
>  on latest sources taken from linux-next will require some time, but of
>  course I will retest and send up-to-date results)
> 
> ---
> Failures: ext4/302 ext4/303 ext4/304 generic/061 generic/063
> generic/075 generic/079 generic/091 generic/112 generic/127
> generic/252 generic/263
> Failed 12 of 200 tests

You didn't say what file system configuration you're using, but I'm
assuming it's a default ext4 4k configuration?  One of the things
about my kvm-xfstests and gce-xfstests setup is that I test a range of
file system configurations, and there are specific exclude files to
skip certain known failures.  Currently we are skipping the following
tests globally (from kvm-xfstests/test-appliances/files/root/fs/ext4/exclude):

# generic/223 tests file alignment, which works on ext4 only by
# accident because we're not RAID stripe aware yet, and works at all
# because we have bias towards aligning on power-of-two block numbers.
# It is a flaky test for some configurations, so skip it.
generic/223

# ext4/304 fails for all configurations, and this appears to be at
# test or fio bug.
#
ext4/304

A recent run of v4.10-rc2 shows that on the 4k configuration, we're only
failing one test:

BEGIN TEST 4k: Ext4 4k block Tue Jan  3 00:20:39 EST 2017
Failures: generic/389

Somewhat older test runs (late in the 4.9 development cycle, before
4.9 final was finally shipped), I saw test failures of generic/082 and
generic/095 as well.

> 1. What is the optimal size for $TEST_DEV ? I see these seconds numbers on a
> small Vm (disk is 16gb):
> 
> ext4/007         22s

With kvm-xfstests and gce-xfstests I normally use a 5G device for
test_dev and scratch_dev except for the bigalloc test configuration
where I use a 20GB device.

> 2.  I see many of the tests are ignored.  Do we have some perfect run,
> reference, Standard, to see how many tests run on up-to-date kernel?
> E.g. I see generic/038 has this output:
>    "[not run] This test requires at least 10GB free".

I currently don't publish one.  What I normally do for myself is run
gce-xfstests at the beginning of the development cycle (e.g., versus
4.10-rc2) and then look for regressions.  The problem is sometimes
regressions are sometimes caused by new tests being added, or changes
pulled in via other trees.

In general the goal is to reduce the number of test failures down to
zero, or else, just suppress them using exclude files, but I haven't
had time to look at a lot of the more recent test failures.  Some of
them, especially for the 1k test case, seem to be quota releated, and
may very be test bugs where the golden output of the tests assume a 4k
block size.  This is definitely true for the bigalloc test cases,
where a number of the failures are known test bugs that we haven't had
time to address.

> Increasing size of $SCRATCH partition will bring the test to live.
> So would be nice to have some reference numbers that those tests
> are expected to run and to pass.

Sure; on my to do list is to update the published kvm-xfstests and
gce-xfstests images to something newer, and when I do that, I can
publish "official" reference test outputs using gce-xfstests.  I've
attached a gce-xfstests output.  It has a huge number of failures
because in the encryption test case because the kernel which was used
for this test run was missing a fix which only landed in Linus's tree
today (see commit fe4f6c801c03b: "fscrypt: fix the
test_dummy_encryption mount option").

You'll see that modulo the encryption failures and some failures on
the 1k config case which I need to track down and iron out, we're in
pretty good shape.  Below please find the summary; attached please
find the full log file.

CMDLINE: full
FSTESTIMG: gce-xfstests/xfstests-201612072310
FSTESTVER: e2fsprogs	v1.43.3-30-g8df85fb (Sun, 4 Sep 2016 21:32:35 -0400)
FSTESTVER: fio		fio-2.14-45-g43f248c (Mon, 24 Oct 2016 20:48:43 -0600)
FSTESTVER: quota		81aca5c (Tue, 12 Jul 2016 16:15:45 +0200)
FSTESTVER: xfsprogs	v4.8.0-rc3 (Mon, 3 Oct 2016 14:25:45 +1100)
FSTESTVER: xfstests-bld	71223ea (Wed, 7 Dec 2016 22:20:14 -0500)
FSTESTVER: xfstests	linux-v3.8-1266-g8d57865 (Wed, 7 Dec 2016 22:56:52 -0500)
FSTESTVER: kernel	4.10.0-rc2-ext4-00001-geb590c0248f4 #176 SMP Tue Jan 3 00:17:59 EST 2017 x86_64
FSTESTCFG: "all"
FSTESTSET: "-g auto"
FSTESTEXC: ""
FSTESTOPT: "aex"
MNTOPTS: ""
CPUS: "2"
MEM: "7477.96"
MEM: 7680 MB (Max capacity)
BEGIN TEST 4k: Ext4 4k block Tue Jan  3 00:20:39 EST 2017
Failures: generic/389
BEGIN TEST 1k: Ext4 1k block Tue Jan  3 01:17:12 EST 2017
Failures: ext4/307 generic/018 generic/076 generic/077 generic/117 generic/233 generic/256 generic/269 generic/270 generic/273 generic/299 generic/300 generic/361 generic/389
BEGIN TEST ext3: Ext4 4k block emulating ext3 Tue Jan  3 02:18:13 EST 2017
Failures: generic/382 generic/389
BEGIN TEST encrypt: Ext4 encryption Tue Jan  3 03:10:27 EST 2017
Failures: ext4/001 ext4/002 ext4/003 ext4/020 ext4/021 ext4/022 ext4/271 ext4/306 ext4/307 ext4/308 generic/002 generic/003 generic/004 generic/005 generic/006 generic/007 generic/008 generic/009 generic/010 generic/012 generic/014 generic/015 generic/016 generic/017 generic/020 generic/021 generic/022 generic/023 generic/024 generic/025 generic/028 generic/029 generic/030 generic/031 generic/032 generic/033 generic/034 generic/035 generic/037 generic/039 generic/040 generic/041 generic/053 generic/056 generic/057 generic/058 generic/059 generic/060 generic/061 generic/062 generic/063 generic/064 generic/065 generic/066 generic/067 generic/069 generic/070 generic/071 generic/073 generic/074 generic/075 generic/077 generic/078 generic/080 generic/084 generic/086 generic/087 generic/088 generic/089 generic/090 generic/098 generic/100 generic/101 generic/102 generic/103 generic/104 generic/105 generic/106 generic/107 generic/109 generic/112 generic/117 generic/120 generic/123 generic/124 generic/126 generic/128 generic/129 generic/132 generic/141 generic/169 generic/177 generic/192 generic/193 generic/213 generic/215 generic/228 generic/234 generic/236 generic/237 generic/241 generic/245 generic/246 generic/247 generic/248 generic/249 generic/255 generic/256 generic/257 generic/274 generic/275 generic/280 generic/286 generic/294 generic/306 generic/307 generic/309 generic/313 generic/316 generic/319 generic/321 generic/322 generic/325 generic/335 generic/336 generic/337 generic/340 generic/341 generic/342 generic/343 generic/344 generic/345 generic/346 generic/354 generic/361 generic/371 generic/375 generic/376 generic/377 generic/378 generic/382 generic/389 generic/391 generic/393 shared/001 shared/272 shared/298
BEGIN TEST nojournal: Ext4 4k block w/ no journal Tue Jan  3 03:28:25 EST 2017
Failures: ext4/301 generic/389
BEGIN TEST ext3conv: Ext4 4k block w/nodelalloc and no flex_bg Tue Jan  3 04:19:52 EST 2017
Failures: generic/347 generic/389
BEGIN TEST adv: Ext4 advanced features (inline_data, metadata_csum, 64bit) Tue Jan  3 05:12:40 EST 2017
Failures: generic/389
BEGIN TEST dioread_nolock: Ext4 4k block w/dioread_nolock Tue Jan  3 06:04:48 EST 2017
Failures: generic/389
BEGIN TEST data_journal: Ext4 4k block w/data=journal Tue Jan  3 06:56:19 EST 2017
Failures: generic/347 generic/389
BEGIN TEST bigalloc: Ext4 4k block w/bigalloc Tue Jan  3 08:12:33 EST 2017
Failures: ext4/004 generic/204 generic/219 generic/235 generic/273 generic/389
BEGIN TEST bigalloc_1k: Ext4 1k block w/bigalloc Tue Jan  3 09:01:49 EST 2017
Failures: ext4/004 generic/204 generic/235 generic/273 generic/389

							- Ted

Download attachment "gce-xfstests.output.xz" of type "application/x-xz" (39760 bytes)

Powered by blists - more mailing lists