From: Luis Henriques <lhenriques@suse.de>
To: Eryu Guan <guan@eryu.me>
Cc: fstests@vger.kernel.org, ceph-devel@vger.kernel.org,
Jeff Layton <jlayton@kernel.org>
Subject: Re: [PATCH] ceph: add a new test for cross quota realms renames
Date: Mon, 23 Nov 2020 09:57:06 +0000 [thread overview]
Message-ID: <87sg90s8el.fsf@suse.de> (raw)
In-Reply-To: <20201122153229.GN3853@desktop> (Eryu Guan's message of "Sun, 22 Nov 2020 23:32:29 +0800")
Eryu Guan <guan@eryu.me> writes:
> On Thu, Nov 19, 2020 at 02:19:56PM +0000, Luis Henriques wrote:
>> For the moment cross quota realms renames has been disabled in CephFS
>> after a bug has been found while renaming files created and truncated.
>> This allowed clients to easily circumvent quotas.
>>
>> Link: https://tracker.ceph.com/issues/48203
>> Signed-off-by: Luis Henriques <lhenriques@suse.de>
>
> Thanks for the test! It'd be great if ceph folks could help review &
> test it as well, as I don't have a cephfs test env yet..
>
> Some minor comments below.
Thanks for your review, Eryu. I'll send out v2 with your comments
implemented.
Cheers,
--
Luis
>> ---
>> tests/ceph/004 | 94 ++++++++++++++++++++++++++++++++++++++++++++++
>> tests/ceph/004.out | 2 +
>> tests/ceph/group | 1 +
>> 3 files changed, 97 insertions(+)
>> create mode 100755 tests/ceph/004
>> create mode 100644 tests/ceph/004.out
>>
>> diff --git a/tests/ceph/004 b/tests/ceph/004
>> new file mode 100755
>> index 000000000000..4021666b138e
>> --- /dev/null
>> +++ b/tests/ceph/004
>> @@ -0,0 +1,94 @@
>> +#! /bin/bash
>> +# SPDX-License-Identifier: GPL-2.0
>> +# Copyright (c) 2020 SUSE Linux Products GmbH. All Rights Reserved.
>> +#
>> +# FS QA Test 004
>> +#
>> +# Tests a bug fix found in cephfs quotas handling. Here's a simplified testcase
>> +# that *should* fail:
>> +#
>> +# mkdir files limit
>> +# truncate files/file -s 10G
>> +# setfattr limit -n ceph.quota.max_bytes -v 1000000
>> +# mv files limit/
>> +#
>> +# Because we're creating a new file and truncating it, we have Fx caps and thus
>> +# the truncate operation will be cached. This prevents the MDSs from updating
>> +# the quota realms and thus the client will allow the above rename(2) to happen.
>> +#
>> +# The bug resulted in dropping support for cross quota-realms renames, reverting
>> +# kernel commit dffdcd71458e ("ceph: allow rename operation under different
>> +# quota realms").
>> +#
>> +# So, the above test will now fail with a -EXDEV or, in the future (when we have
>> +# a proper fix), with -EDQUOT.
>> +#
>> +# This bug was tracker here:
>> +#
>> +# https://tracker.ceph.com/issues/48203
>> +#
>> +seq=`basename $0`
>> +seqres=$RESULT_DIR/$seq
>> +echo "QA output created by $seq"
>> +
>> +here=`pwd`
>> +tmp=/tmp/$$
>> +status=1 # failure is the default!
>> +trap "_cleanup; exit \$status" 0 1 2 3 15
>> +
>> +_cleanup()
>> +{
>> + cd /
>> + rm -f $tmp.*
>> +}
>> +
>> +# get standard environment, filters and checks
>> +. ./common/rc
>> +. ./common/filter
>> +. ./common/attr
>> +
>> +# remove previous $seqres.full before test
>> +rm -f $seqres.full
>> +
>> +# real QA test starts here
>> +
>> +_supported_fs ceph
>> +_require_attrs
>> +_require_test
>
> Need '_require_test_program "rename"' as well.
>
>> +
>> +workdir=$TEST_DIR/test-$seq
>> +
>> +orig1=$workdir/orig1
>> +orig2=$workdir/orig2
>> +file1=$orig1/file
>> +file2=$orig2/file
>> +dest=$workdir/dest
>> +
>> +rm -rf $workdir
>> +mkdir $workdir
>> +mkdir $orig1 $orig2 $dest
>> +
>> +# set quota to 1m
>> +$SETFATTR_PROG -n ceph.quota.max_bytes -v 1000000 $dest
>> +# set quota to 20g
>> +$SETFATTR_PROG -n ceph.quota.max_bytes -v 20000000000 $orig2
>> +
>> +#
>> +# The following 2 testcases shall fail with either -EXDEV or -EDQUOT
>> +#
>> +
>> +# from 'root' realm to $dest realm
>> +$XFS_IO_PROG -f -c "truncate 10G" $file1
>> +$here/src/rename $orig1 $dest/new1 >> $seqres.full 2>&1
>> +[ $? -ne 1 ] && _fatal "cross quota realms rename succeeded"
>
> Use _fail instead of _fatal in test, _fatal is usually used in common
> helpers, to report internal errors such as wrong usage of helpers.
>
>> +
>> +# from $orig2 realm to $dest realm
>> +$XFS_IO_PROG -f -c "truncate 10G" $file2
>> +$here/src/rename $orig2 $dest/new2 >> $seqres.full 2>&1
>> +[ $? -ne 1 ] && _fatal "cross quota realms rename succeeded"
>
> Same here.
>
> Thanks,
> Eryu
>
>> +
>> +echo "Silence is golden"
>> +
>> +# success, all done
>> +status=0
>> +exit
>> diff --git a/tests/ceph/004.out b/tests/ceph/004.out
>> new file mode 100644
>> index 000000000000..af8614ae45ac
>> --- /dev/null
>> +++ b/tests/ceph/004.out
>> @@ -0,0 +1,2 @@
>> +QA output created by 004
>> +Silence is golden
>> diff --git a/tests/ceph/group b/tests/ceph/group
>> index adbf61547766..47903d21966c 100644
>> --- a/tests/ceph/group
>> +++ b/tests/ceph/group
>> @@ -1,3 +1,4 @@
>> 001 auto quick copy
>> 002 auto quick copy
>> 003 auto quick copy
>> +004 auto quick quota
next prev parent reply other threads:[~2020-11-23 9:57 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-19 14:19 [PATCH] ceph: add a new test for cross quota realms renames Luis Henriques
2020-11-22 15:32 ` Eryu Guan
2020-11-23 9:57 ` Luis Henriques [this message]
2020-11-23 10:34 ` [PATCH v2] " Luis Henriques
2020-11-23 12:28 ` Jeff Layton
2020-11-23 14:43 ` Luis Henriques
2020-11-23 15:39 ` Jeff Layton
2020-11-23 16:24 ` Luis Henriques
2020-11-23 16:39 ` Jeff Layton
2020-11-23 17:25 ` Luis Henriques
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=87sg90s8el.fsf@suse.de \
--to=lhenriques@suse.de \
--cc=ceph-devel@vger.kernel.org \
--cc=fstests@vger.kernel.org \
--cc=guan@eryu.me \
--cc=jlayton@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.