Re: [PATCH v2] ceph/005: verify correct statfs behaviour with quotas

[Xiubo Li <xiubli@redhat.com>]

On 6/27/22 5:20 PM, Luís Henriques wrote:
> On Mon, Jun 27, 2022 at 08:35:14AM +0800, Xiubo Li wrote:
>> Hi Luis,
>>
>> Sorry for late.
>>
>> On 6/15/22 11:14 PM, Luís Henriques wrote:
>>> When using a directory with 'max_bytes' quota as a base for a mount,
>>> statfs shall use that 'max_bytes' value as the total disk size.  That
>>> value shall be used even when using subdirectory as base for the mount.
>>>
>>> A bug was found where, when this subdirectory also had a 'max_files'
>>> quota, the real filesystem size would be returned instead of the parent
>>> 'max_bytes' quota value.  This test case verifies this bug is fixed.
>>>
>>> Signed-off-by: Luís Henriques <lhenriques@suse.de>
>>> ---
>>> Finally, I've managed to come back to this test.  The major changes since
>>> v1 are:
>>>    - creation of an helper for getting total mount space using 'df'
>>>    - now the test sends quota size to stdout
>>>
>>>    common/rc          | 13 +++++++++++++
>>>    tests/ceph/005     | 38 ++++++++++++++++++++++++++++++++++++++
>>>    tests/ceph/005.out |  4 ++++
>>>    3 files changed, 55 insertions(+)
>>>    create mode 100755 tests/ceph/005
>>>    create mode 100644 tests/ceph/005.out
>>>
>>> diff --git a/common/rc b/common/rc
>>> index 2f31ca464621..72eabb7a428c 100644
>>> --- a/common/rc
>>> +++ b/common/rc
>>> @@ -4254,6 +4254,19 @@ _get_available_space()
>>>    	echo $((avail_kb * 1024))
>>>    }
>>> +# get the total space in bytes
>>> +#
>>> +_get_total_space()
>>> +{
>>> +	if [ -z "$1" ]; then
>>> +		echo "Usage: _get_total_space <mnt>"
>>> +		exit 1
>>> +	fi
>>> +	local total_kb;
>>> +	total_kb=`$DF_PROG $1 | tail -n1 | awk '{ print $3 }'`
>>> +	echo $(($total_kb * 1024))
>>> +}
>>> +
>>>    # return device size in kb
>>>    _get_device_size()
>>>    {
>>> diff --git a/tests/ceph/005 b/tests/ceph/005
>>> new file mode 100755
>>> index 000000000000..7eb687e8a092
>>> --- /dev/null
>>> +++ b/tests/ceph/005
>>> @@ -0,0 +1,38 @@
>>> +#! /bin/bash
>>> +# SPDX-License-Identifier: GPL-2.0
>>> +# Copyright (C) 2022 SUSE Linux Products GmbH. All Rights Reserved.
>>> +#
>>> +# FS QA Test 005
>>> +#
>>> +# Make sure statfs reports correct total size when:
>>> +# 1. using a directory with 'max_byte' quota as base for a mount
>>> +# 2. using a subdirectory of the above directory with 'max_files' quota
>>> +#
>>> +. ./common/preamble
>>> +_begin_fstest auto quick quota
>>> +
>>> +_supported_fs ceph
>>> +_require_scratch
>>> +
>>> +_scratch_mount
>>> +mkdir -p "$SCRATCH_MNT/quota-dir/subdir"
>>> +
>>> +# set quota
>>> +quota=$((2 ** 30)) # 1G
>>> +$SETFATTR_PROG -n ceph.quota.max_bytes -v "$quota" "$SCRATCH_MNT/quota-dir"
>>> +$SETFATTR_PROG -n ceph.quota.max_files -v "$quota" "$SCRATCH_MNT/quota-dir/subdir"
>>> +_scratch_unmount
>>> +
>>> +SCRATCH_DEV="$SCRATCH_DEV/quota-dir" _scratch_mount
>>> +echo ceph quota size is $(_get_total_space "$SCRATCH_MNT") bytes
>>> +SCRATCH_DEV="$SCRATCH_DEV/quota-dir" _scratch_unmount
>>> +
>> For the 'SCRATCH_DEV' here, if you do:
>>
>> SCRATCH_DEV="$SCRATCH_DEV/quota-dir"
>>
>> twice, won't it be "${ceph_scratch_dev}/quota-dir/quota-dir" at last ?
>>
>> Shouldn't it be:
>>
>> SCRATCH_DEV="$TEST_DIR/quota-dir" ?
> No, actually we do really need to have $SCRATCH_DEV and not $TEST_DIR
> here, because we want to have SCRATCH_DEV set to something like:
>
> 	<mon-ip-addr>:<port>:/quota-dir
>
> so that we will use that directory (with quotas) as base for the mount.
>
> Regarding the second attribution using the already modifed $SCRATCH_DEV
> variable, that's not really what is happening (and I had to go
> double-check myself, as you got me confused too :-).
>
> So, if you do:
>
> SCRATCH_DEV="$SCRATCH_DEV/quota-dir" _scratch_mount
> SCRATCH_DEV="$SCRATCH_DEV/quota-dir" _scratch_unmount
>
> the SCRATCH_DEV value is changed *only* for the _scratch_[un]mount
> functions, but SCRATCH_DEV value isn't really changed after these
> attributions.  This is probably a bashism (I'm not really sure), but you
> can see a similar pattern in other places (see, for example, test
> xfs/234).

Sorry, could you give the link about this ?

Checked the xfs/234, I didn't find any place is using the similar pattern.

This is what I see:

  53 # Now restore the obfuscated one back and take a look around
  54 echo "Restore metadump"
  55 xfs_mdrestore $metadump_file $TEST_DIR/image
  56 SCRATCH_DEV=$TEST_DIR/image _scratch_mount
  57 SCRATCH_DEV=$TEST_DIR/image _scratch_unmount
  58
  59 echo "Check restored fs"
  60 _check_generic_filesystem $metadump_file



> Anyway, if you prefer, I'm fine sending v3 of this test doing something
> like:

Locally I just test this use case, it seems working as my guess.

-- Xiubo

>
> SCRATCH_DEV_ORIG="$SCRATCH_DEV"
> SCRATCH_DEV="$SCRATCH_DEV_ORIG/quota-dir" _scratch_mount
> ...
>
> Cheers,
> --
> Luís
>
>> -- Xiubo
>>
>>> +SCRATCH_DEV="$SCRATCH_DEV/quota-dir/subdir" _scratch_mount
>>> +echo subdir ceph quota size is $(_get_total_space "$SCRATCH_MNT") bytes
>>> +SCRATCH_DEV="$SCRATCH_DEV/quota-dir/subdir" _scratch_unmount
>>> +
>>> +echo "Silence is golden"
>>> +
>>> +# success, all done
>>> +status=0
>>> +exit
>>> diff --git a/tests/ceph/005.out b/tests/ceph/005.out
>>> new file mode 100644
>>> index 000000000000..47798b1fcd6f
>>> --- /dev/null
>>> +++ b/tests/ceph/005.out
>>> @@ -0,0 +1,4 @@
>>> +QA output created by 005
>>> +ceph quota size is 1073741824 bytes
>>> +subdir ceph quota size is 1073741824 bytes
>>> +Silence is golden
>>>