From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Eryu Guan <guaneryu@gmail.com>
Cc: Chandan Rajendra <chandanrlinux@gmail.com>,
fstests@vger.kernel.org, chandan@linux.ibm.com,
hch@infradead.org, linux-xfs@vger.kernel.org
Subject: Re: [PATCH] common/xfs: Execute _xfs_check only when explicitly asked
Date: Mon, 6 Apr 2020 22:48:22 -0700 [thread overview]
Message-ID: <20200407054822.GA6740@magnolia> (raw)
In-Reply-To: <20200406123030.GG3128153@desktop>
On Mon, Apr 06, 2020 at 08:31:07PM +0800, Eryu Guan wrote:
> On Mon, Mar 30, 2020 at 03:42:03PM +0530, Chandan Rajendra wrote:
> > fsstress when executed as part of some of the tests (e.g. generic/270)
> > invokes chown() syscall many times by passing random integers as value
> > for the uid argument. For each such syscall invocation for which there
> > is no on-disk quota block, xfs invokes xfs_dquot_disk_alloc() which
> > allocates a new block and instantiates all the quota structures mapped
> > by the newly allocated block. For filesystems with 64k block size, the
> > number of on-disk quota structures created will be 16 times more than
> > that for a filesystem with 4k block size.
> >
> > xfs_db's check command (executed after test script finishes execution)
> > will try to read in all of the on-disk quota structures into
> > memory. This causes the OOM event to be triggered when reading from
> > filesystems with 64k block size. For filesystems with sufficiently large
> > amount of system memory, this causes the test to execute for a very long
> > time.
> >
> > Due to the above stated reasons, this commit disables execution of
> > xfs_db's check command unless explictly asked by the user by setting
> > $EXECUTE_XFS_DB_CHECK variable.
> >
> > Signed-off-by: Chandan Rajendra <chandanrlinux@gmail.com>
> > ---
> > README | 2 ++
> > common/xfs | 17 +++++++++++++----
> > 2 files changed, 15 insertions(+), 4 deletions(-)
> >
> > diff --git a/README b/README
> > index 094a7742..c47569cd 100644
> > --- a/README
> > +++ b/README
> > @@ -88,6 +88,8 @@ Preparing system for tests:
> > run xfs_repair -n to check the filesystem; xfs_repair to rebuild
> > metadata indexes; and xfs_repair -n (a third time) to check the
> > results of the rebuilding.
> > + - set EXECUTE_XFS_DB_CHECK=1 to have _check_xfs_filesystem
> > + run xfs_db's check command on the filesystem.
>
> It seems spaces are used for indention instead of tab in README.
>
> > - xfs_scrub, if present, will always check the test and scratch
> > filesystems if they are still online at the end of the test.
> > It is no longer necessary to set TEST_XFS_SCRUB.
> > diff --git a/common/xfs b/common/xfs
> > index d9a9784f..93ebab75 100644
> > --- a/common/xfs
> > +++ b/common/xfs
> > @@ -455,10 +455,19 @@ _check_xfs_filesystem()
> > ok=0
> > fi
> >
> > - # xfs_check runs out of memory on large files, so even providing the test
> > - # option (-t) to avoid indexing the free space trees doesn't make it pass on
> > - # large filesystems. Avoid it.
> > - if [ "$LARGE_SCRATCH_DEV" != yes ]; then
> > + dbsize="$($XFS_INFO_PROG "${device}" | grep data.*bsize | sed -e 's/^.*bsize=//g' -e 's/\([0-9]*\).*$/\1/g')"
>
> This is probably a left-over from v1 patch, which is not needed in v2.
>
> > +
> > + # xfs_check runs out of memory,
> > + # 1. On large files. So even providing the test option (-t) to
> > + # avoid indexing the free space trees doesn't make it pass on
> > + # large filesystems.
> > + # 2. When checking filesystems with large number of quota
> > + # structures. This case happens consistently with 64k blocksize when
> > + # creating large number of on-disk quota structures whose quota ids
> > + # are spread across a large integer range.
> > + #
> > + # Hence avoid executing it unless explicitly asked by user.
> > + if [ -n "$EXECUTE_XFS_DB_CHECK" -a "$LARGE_SCRATCH_DEV" != yes ]; then
> > _xfs_check $extra_log_options $device 2>&1 > $tmp.fs_check
>
> Looks fine to me, I'd like to see xfs_check being disabled. But it'd be
> great if xfs folks could ack it as well.
I think I'd rather we just do the testing to see what things xfs_check
catches that xfs_repair doesn't, and then we can deprecate running
xfs_check in fstests (by default) entirely. In the long run it doesn't
make sense to maintain /three/ separate metadata verification utilities.
--D
> Thanks,
> Eryu
>
> > fi
> > if [ -s $tmp.fs_check ]; then
> > --
> > 2.19.1
> >
prev parent reply other threads:[~2020-04-07 5:52 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-30 10:12 [PATCH] common/xfs: Execute _xfs_check only when explicitly asked Chandan Rajendra
2020-04-06 12:31 ` Eryu Guan
2020-04-07 5:48 ` Darrick J. Wong [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200407054822.GA6740@magnolia \
--to=darrick.wong@oracle.com \
--cc=chandan@linux.ibm.com \
--cc=chandanrlinux@gmail.com \
--cc=fstests@vger.kernel.org \
--cc=guaneryu@gmail.com \
--cc=hch@infradead.org \
--cc=linux-xfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox