[sles-beta] btrfs & tooling
Jeff Mahoney
jeffm at suse.com
Sat Sep 27 09:55:49 MDT 2014
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 9/24/14, 7:19 AM, Christoph Schumacher wrote:
> Hi all,
Hi Christoph -
> I lately tested RC2 with focus btrfs, subvolumes and tooling.
>
> I was disappointed to see, that there still are a load of issues
> open to especially quota and tooling with btrfs as i mentioned
> before. - grp-quota seems to not work at all (i already discussed
> with our SE)
I'm not sure what issues you're referring to here. Can you elaborate?
> - human readable quota-size list does not exist
Are you referring to the btrfs qgroup show output? Yes, this could be
better. While the output presents all the information correctly, it's
not easily translatable by the user. It would probably be more helpful
to map the qgroups to subvolumes where possible. As far as the columns
go, they correspond to the number of bytes referenced by a subvolume
(rfer) and the number of bytes referenced exclusively by a subvolume
(excl).
> - df ....still no hooks for btrfs quota/subvolumes
How would this information be presented? There is no direct mapping in
the way there is for LVM volumes. All subvolumes share a common
storage pool unless otherwise restricted. There's no way to determine
how much space a subvolume occupies because the space referenced by it
can be a mix of shared references and exclusive references. So we
could present 'used' on a subvolume basis, but if you added up the
'used' space across all subvolumes, we'd end up with a number that
could be much larger than the size of the file system.
> ...at this point my testing stoped...
>
> As long as the only option is to just create an entire diskvolume
> (without lvm) because of the bad subvolume handling, i see btrfs
> not in production stage. As soon as lvm enters the stage, i see no
> gain in btrfs.
Btrfs subvolumes and LVM subvolumes are different and this is by
design. Sure, you can make a loose analogy by using dm thin with
snapshots, but the usage/mapping knowledge is still at the block
layer. Btrfs subvolumes offer dedupe, reflink copies, and a number of
other space-saving operations.
- -Jeff
>
> I'll place all my concerns next week at the expert forum btrfs
> workshop, but i had the feeling to point this out on the list. What
> about all list participants? Do you plan to use btrfs for
> production (at least the OS part)?
>
> *Dipl.-Inf. Christoph Schumacher* FAI5 Competencecenter Dezentrale
> Server Dr. Ing. h.c. F. Porsche AG Porscheplatz 1, D-70435
> Stuttgart Zuffenhausen Telefon: +49 (0)711/911-23729 Fax:
> +49 (0)711/911-23171 E-Mail:
> christoph.schumacher at porsche.de
>
>
>
> From: Christoph Schumacher
> <Christoph.Schumacher at porsche.de> To:
> sles-beta at lists.suse.com Date: 28/05/2014 14:00 Subject:
> Re: [sles-beta] btrfs & tooling Sent by:
> sles-beta-bounces at lists.suse.com
> ------------------------------------------------------------------------
>
>
>
>
> Hi Richard,
>
> Honestly it'll be a mess to dismiss "df" for a lot of people,
> especially the guys coming from any UNIX.
>
> As i mentioned: on my beta7 install the commands you issued do not
> work well or hard to read:
>
> For example btrfs fi df shows a lot of information and it's hard
> to filter at a glance.
>
> btrfs qgroup show does not work for me at all....what's the quota?
> have a guess!
>
> # btrfs qgroup show /tmp qgroupid rfer excl -------- ----
> ---- 0/5 16384 16384 0/257 2359185408 2359185408 0/258
> 16384 16384 0/259 102576128 102576128 0/260 16384
> 16384 0/261 16384 16384 0/262 16384 16384 0/263
> 97067008 97067008
>
>
> * Dipl.-Inf. Christoph Schumacher* FAI5 Competence Center
> Dezentrale Server Dr. Ing. h.c. F. Porsche AG Porscheplatz 1,
> D-70435 Stuttgart Zuffenhausen Telefon: +49 (0)711/911-23729 Fax:
> +49 (0)711/911-23171 E-Mail:
> christoph.schumacher at porsche.de
>
>
>
> From: Richard Brown <rbrown at suse.de> To:
> sles-beta at lists.suse.com Date: 28/05/2014 13:53 Subject:
> Re: [sles-beta] btrfs & tooling Sent by:
> sles-beta-bounces at lists.suse.com
> ------------------------------------------------------------------------
>
>
>
>
> Hello Christoph,
>
> I'm afraid I can't answer all your questions, but I'll take a stab
> at a few that I believe I know the answer to
>
> On Wed, 2014-05-28 at 13:39 +0200, Christoph Schumacher wrote:
>> - on subvolume with limits/quota df always shows the size of the
>> parent volume also for all subvolumes IMHO it should either show
>> the limited size or not show the subvolume at all because it's
>> only disturbing
>
>> - also after "mirroring" a btrfs with a raid1 df shows the sum of
>> both volumes in the btrfs despite the real size
>
> Because of the nature of how btrfs works, df isn't the best tool to
> use
>
> btrfs has it's own tools for the purpose, such as 'btrfs fi show'
> which shows you the raw disk usage, and 'btrfs fi df <filepath>'
> which will show you the equivalent of df, taking into account
> things like btrfs RAID, for the subvolume selected by <filepath>
>
>> - i can't figure out how to show the quota of a btrfs volume
>> easily and clearly ..... what the common known btrfs-tools show
>> is cr....not good.
>
> 'btrfs qgroup show <filepath>' where <filepath> is a location for
> your btrfs volume/subvolume, should give you the information you
> require.
>
> Hope this helps,
>
> - Rich
>
> --
> -------------------------------------------------------------------
>
>
Richard Brown, QA Engineer
> Phone +4991174053-361, Fax +4991174053-483 SUSE LINUX Products
> GmbH, Maxfeldstr. 5, D-90409 Nuernberg Geschaeftsfuehrer: Jeff
> Hawn, Jennifer Guild, Felix Imendoerffer, HRB 16746 (AG Nuernberg)
> -------------------------------------------------------------------
>
>
>
> _______________________________________________ sles-beta mailing
> list sles-beta at lists.suse.com_
> __http://lists.suse.com/mailman/listinfo/sles-beta_
>
>
> Dr. Ing. h.c. F. Porsche Aktiengesellschaft Sitz der Gesellschaft:
> Stuttgart Registergericht: Amtsgericht Stuttgart HRB-Nr. 730623
> Vorsitzender des Aufsichtsrats: Dr. Wolfgang Porsche Vorstand:
> Matthias Müller, Vorsitzender Thomas Edig, stv. Vorsitzender Dr.
> Oliver Blume, Wolfgang Hatz, Bernhard Maier, Lutz Meschke,
> Uwe-Karsten Städter
>
> Die vorgenannten Angaben werden jeder E-Mail automatisch
> hinzugefügt. Dies ist kein Anerkenntnis, dass es sich beim Inhalt
> dieser E-Mail um eine rechtsverbindliche Erklärung der Porsche AG
> handelt. Erklärungen, die die Porsche AG verpflichten, bedürfen
> jeweils der Unterschrift durch zwei zeichnungs- berechtigte
> Personen der AG._______________________________________________
> sles-beta mailing list sles-beta at lists.suse.com
> http://lists.suse.com/mailman/listinfo/sles-beta
>
>
> Dr. Ing. h.c. F. Porsche Aktiengesellschaft Sitz der Gesellschaft:
> Stuttgart Registergericht: Amtsgericht Stuttgart HRB-Nr. 730623
> Vorsitzender des Aufsichtsrats: Dr. Wolfgang Porsche Vorstand:
> Matthias Müller, Vorsitzender Thomas Edig, stv. Vorsitzender Dr.
> Oliver Blume, Wolfgang Hatz, Bernhard Maier, Lutz Meschke,
> Uwe-Karsten Städter
>
> Die vorgenannten Angaben werden jeder E-Mail automatisch
> hinzugefügt. Dies ist kein Anerkenntnis, dass es sich beim Inhalt
> dieser E-Mail um eine rechtsverbindliche Erklärung der Porsche AG
> handelt. Erklärungen, die die Porsche AG verpflichten, bedürfen
> jeweils der Unterschrift durch zwei zeichnungs- berechtigte
> Personen der AG.
>
>
> _______________________________________________ sles-beta mailing
> list sles-beta at lists.suse.com
> http://lists.suse.com/mailman/listinfo/sles-beta
>
- --
Jeff Mahoney
SUSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2.0.19 (Darwin)
iQIcBAEBAgAGBQJUJt4FAAoJEB57S2MheeWy1MsQAMQIWLyfRzrgloR0Gt21rdiF
HTtpfgBbsrdxzHWBG9Z2YVBQdjTdC5mQMz81/2aUnqsOX+LO5PznMlShP+3VykUv
qXCRxH42D0zKbID2GwAgX7Y3446kmMfHJu70DRgkQMy1n+WzAl7dWXe2dqyCrKlw
V+oXto2qZjWJPJ83HKqqDSDDjzCExM4bLVbamdT3O2sLs6fDSM2R8OBA61RdBEwp
1lTQ5ePwIm5E7DyztwyOIN5FutMzKVHyIOq5B5QxQAovQdXJXsl5xfvh7daknTmf
oEvpmx2fzJn7qIi3oB+C1PkpcJeSFFJa0HtemRvVRXtUzyiLHkPdIBMAe1NxRjr9
k7io6kvEOTP/CXsHJ8Nuu0r1zev/HvtF3+u6hK15GdNp2qTtFvAdgCGmo3uUwSmF
9CdllKUJgk5SLaNHqf47mXT/2WbX8ygBZA5Dp1EewRJWLOOKz0Kd61ehrm61xdhY
9EfUJAfVu0Hz/WbHuZeAqVRdIQhd82qfoXN1SX1R1rW82LJFK7IVZqGTpLS1EIvC
5Zo3v+bag6ugxi8fCwSmSvfUkQAS/guMMn+1P4Co5pyzLByPM6Zzy2F+g5HoZ52g
Gy3ZbMH0gI01T59E3RQCmvc9MmwV+jnshyN3MyY41HopfcjRA4f1cxcib4mgYypJ
8n5Cbx9TUdgQZSWwqev7
=S8xi
-----END PGP SIGNATURE-----
More information about the sles-beta
mailing list