Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Robin Bowes]

On Fri, June 4, 2004 7:29, Neil Brown said:
>   - Further support for partitionable arrays included "--auto=" option
>     and "auto=" config file entry which instructs mdadm to create the necessary
>     device files after allocating an unused array number.

Neil,

I am interpreting this to mean that I can create, for example, a large (e.g. 1TB) RAID5
array and create smaller partitions (/home, /usr, etc) on top of the RAID5 array? Is
this correct?

Where can I read more about this?

Thanks,

R.
-- 
http://robinbowes.com

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Andreas Haumer]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi!

Robin Bowes wrote:
> On Fri, June 4, 2004 7:29, Neil Brown said:
>
>>  - Further support for partitionable arrays included "--auto=" option
>>    and "auto=" config file entry which instructs mdadm to create the necessary
>>    device files after allocating an unused array number.
>
>
> Neil,
>
> I am interpreting this to mean that I can create, for example, a large (e.g. 1TB) RAID5
> array and create smaller partitions (/home, /usr, etc) on top of the RAID5 array? Is
> this correct?
>
Hm.
Why would anyone use this, given that we can put LVM on
top of SW-RAID?

root@tolstoi:~ {593} $ vgdisplay -v
- --- Volume group ---
VG Name               sys
VG Access             read/write
VG Status             available/resizable
VG #                  0
MAX LV                256
Cur LV                8
Open LV               8
MAX LV Size           2 TB
Max PV                256
Cur PV                1
Act PV                1
VG Size               136.66 GB
PE Size               32 MB
Total PE              4373
Alloc PE / Size       4170 / 130.31 GB
Free  PE / Size       203 / 6.34 GB
VG UUID               4I7XyX-nuyy-gnJY-XciQ-hlu7-Yr03-r7eoD8

- --- Logical volume ---
LV Name                /dev/sys/swap1
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   1
# open                 1
LV Size                2 GB
Current LE             64
Allocated LE           64
Allocation             next free
Read ahead sectors     1024
Block device           58:0

- --- Logical volume ---
LV Name                /dev/sys/swap2
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   2
# open                 1
LV Size                2 GB
Current LE             64
Allocated LE           64
Allocation             next free
Read ahead sectors     1024
Block device           58:1

- --- Logical volume ---
LV Name                /dev/sys/root
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   3
# open                 1
LV Size                320 MB
Current LE             10
Allocated LE           10
Allocation             next free
Read ahead sectors     1024
Block device           58:2

- --- Logical volume ---
LV Name                /dev/sys/usr
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   4
# open                 1
LV Size                4 GB
Current LE             128
Allocated LE           128
Allocation             next free
Read ahead sectors     1024
Block device           58:3

- --- Logical volume ---
LV Name                /dev/sys/opt
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   5
# open                 1
LV Size                4 GB
Current LE             128
Allocated LE           128
Allocation             next free
Read ahead sectors     1024
Block device           58:4

- --- Logical volume ---
LV Name                /dev/sys/var
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   6
# open                 1
LV Size                4 GB
Current LE             128
Allocated LE           128
Allocation             next free
Read ahead sectors     1024
Block device           58:5

- --- Logical volume ---
LV Name                /dev/sys/tmp
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   7
# open                 1
LV Size                4 GB
Current LE             128
Allocated LE           128
Allocation             next free
Read ahead sectors     1024
Block device           58:6

- --- Logical volume ---
LV Name                /dev/sys/work
VG Name                sys
LV Write Access        read/write
LV Status              available
LV #                   8
# open                 1
LV Size                110 GB
Current LE             3520
Allocated LE           3520
Allocation             next free
Read ahead sectors     1024
Block device           58:7


- --- Physical volumes ---
PV Name (#)           /dev/md/0 (1)
PV Status             available / allocatable
Total PE / Free PE    4373 / 203

root@tolstoi:~ {595} $ mdadm --detail /dev/md/0
/dev/md/0:
        Version : 00.90.00
  Creation Time : Sun Feb 22 19:45:15 2004
     Raid Level : raid5
     Array Size : 143331712 (136.69 GiB 146.77 GB)
    Device Size : 71665856 (68.35 GiB 73.39 GB)
   Raid Devices : 3
  Total Devices : 3
Preferred Minor : 0
    Persistence : Superblock is persistent

    Update Time : Thu Jun  3 16:52:42 2004
          State : dirty, no-errors
 Active Devices : 3
Working Devices : 3
 Failed Devices : 0
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 64K

    Number   Major   Minor   RaidDevice State
       0       8        2        0      active sync   /dev/scsi/host0/bus0/target0/lun0/part2
       1       8       18        1      active sync   /dev/scsi/host0/bus0/target1/lun0/part2
       2       8       34        2      active sync   /dev/scsi/host0/bus0/target2/lun0/part2
           UUID : 7abbd703:643ffe8e:d6ae4d0e:1b0908aa
         Events : 0.42

root@tolstoi:~ {597} $ mount
rootfs on / type rootfs (ro)
/dev/sys/root on / type reiserfs (ro)
devfs on /dev type devfs (rw)
proc on /proc type proc (rw)
/dev/sys/usr on /usr type reiserfs (ro)
/dev/sys/opt on /opt type reiserfs (ro)
/dev/sys/var on /var type reiserfs (rw,noexec)
/dev/sys/tmp on /tmp type reiserfs (rw,noexec)
/dev/sys/work on /work type reiserfs (rw,noexec)
/dev/scsi/host0/bus0/target0/lun0/part1 on /boot type ext2 (ro)
tmpfs on /dev/shm type tmpfs (rw,noexec)


What are the advantages of a partitionable md device
compared to a SWRAID+LVM configuration?

- - andreas

- --
Andreas Haumer                     | mailto:andreas@xss.co.at
*x Software + Systeme              | http://www.xss.co.at/
Karmarschgasse 51/2/20             | Tel: +43-1-6060114-0
A-1100 Vienna, Austria             | Fax: +43-1-6060114-71
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFAw0aWxJmyeGcXPhERAqc8AJwOJygJV6tL33Sswj5I7nKfm6IwWQCggoA4
A5GHH+GhDYHDqFBzCpm9ElA=
=q/Rp
-----END PGP SIGNATURE-----

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Neil Brown]

On Sunday June 6, andreas@xss.co.at wrote:
> 
> Hi!
> 
> Robin Bowes wrote:
> > Neil,
> >
> > I am interpreting this to mean that I can create, for example, a large (e.g. 1TB) RAID5
> > array and create smaller partitions (/home, /usr, etc) on top of the RAID5 array? Is
> > this correct?
> >
> Hm.
> Why would anyone use this, given that we can put LVM on
> top of SW-RAID?

> 
> 
> What are the advantages of a partitionable md device
> compared to a SWRAID+LVM configuration?
> 

Familiarity.  Simplicity.  Uniformity with other partitioning.

LVM certainly works, now "traditional" partitioning does too...

Choice is good.  Use whichever suits you.

NeilBrown

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Mario 'BitKoenig' Holbe]

Andreas Haumer <andreas@xss.co.at> wrote:
> What are the advantages of a partitionable md device
> compared to a SWRAID+LVM configuration?

You don't need LVM and thus you have one less layer
with potential failures and errors.


regards,
   Mario
-- 
<delta> talk softly and carry a keen sword

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Joe Pruett]

> You don't need LVM and thus you have one less layer
> with potential failures and errors.

using lvm for just partitioning would be silly.  but lvm gives you much
more than that.  the ability to span partitions over multiple raid arrays.  
and the main reason i use it is to be able to grow/shrink storage if
something changes in the future.

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Mario 'BitKoenig' Holbe]

Joe Pruett <joey@clean.q7.com> wrote:
> using lvm for just partitioning would be silly.  but lvm gives you much

You got the point :)

> more than that.  the ability to span partitions over multiple raid arrays.  

And if you don't need or want all that, it would
be silly to use LVM - as you mentioned above :)
That's why it makes sense to be able to partition
RAIDs without the need of LVM.


regards,
   Mario
-- 
I thought the only thing the internet was good for was porn.  -- Futurama

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Clemens Schwaighofer]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Andreas Haumer wrote:
| Hi!
|
| Robin Bowes wrote:
|
|>>On Fri, June 4, 2004 7:29, Neil Brown said:
|>>
|>>
|>>> - Further support for partitionable arrays included "--auto=" option
|>>>   and "auto=" config file entry which instructs mdadm to create the
necessary
|>>>   device files after allocating an unused array number.
|>>
|>>
|>>Neil,
|>>
|>>I am interpreting this to mean that I can create, for example, a
large (e.g. 1TB) RAID5
|>>array and create smaller partitions (/home, /usr, etc) on top of the
RAID5 array? Is
|>>this correct?
|>>
|
| Hm.
| Why would anyone use this, given that we can put LVM on
| top of SW-RAID?

Well I'd say to leave out one area of problems. LVM might get buggy or
cause troubles. And if you can create partitions in a SW raid you can
avoid that.

eg on a firewall I want only real necessary software and LVM is
definitly not one of them ...

- --
Clemens Schwaighofer - IT Engineer & System Administration
==========================================================
TEQUILA\Japan, 6-17-2 Ginza Chuo-ku, Tokyo 104-8167, JAPAN
Tel: +81-(0)3-3545-7703            Fax: +81-(0)3-3545-7343
http://www.tequila.co.jp
==========================================================
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAxrXWjBz/yQjBxz8RAnu/AJ0YlAnEXGxK3hBiGmSWJiOfT3O+rQCgheXZ
V5DTW6ESA7M8W9ZbUklefd8=
=jwFK
-----END PGP SIGNATURE-----

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Norman Schmidt]

Neil Brown schrieb:

> The simplest way to access this functionality is with
> 
>     mdadm --create /dev/md/XX --auto=partition --level= ....

I finally got it. After mkdir /dev/md, that works.

One more thing: It would be helpful to have a reference between the device 
name (which is user for mounting, partitionimg etc) and the /proc/mdstat 
entry. In my case that would be /dev/md/mdp1 and md_d0.

For unpartitionable raid this is no problem, since the device name occurs 
in /proc/mdstat directly.

Thanks for the great work so far.

Norman Schmidt.

ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Neil Brown]

I am pleased to announce the availability of 
   mdadm version 1.6.0
It is available at
   http://www.cse.unsw.edu.au/~neilb/source/mdadm/
and
   http://www.{countrycode}.kernel.org/pub/linux/utils/raid/mdadm/

as a source tar-ball and (at the first site) as an SRPM, and as an RPM for i386.

mdadm is a tool for creating, managing and monitoring
device arrays using the "md" driver in Linux, also
known as Software RAID arrays.

Release 1.6.0 adds:
  - --grow which (in 2.6.7-rc1-mm1 and hopefully 2.6.8) allows raid1/4/5/6
    arrays to change the active size of the underlying devices, and allows
    raid1 arrays to change the number of active drives.
  - Allows --build to buld raid1 and multipath arrays.
  - adds "degraded" and "recovering" as possibilities for the status line
    in --detail
  - fixes a bug in 1.5.0 which stopped resync status messages from being
    generated in --monitor mode
  - Further support for partitionable arrays included "--auto=" option
    and "auto=" config file entry which instructs mdadm to create the necessary
    device files after allocating an unused array number.
  - assorted minor fixes and improvements.

Development of mdadm is sponsored by CSE@UNSW: 
  The School of Computer Science and Engineering
at
  The University of New South Wales

NeilBrown  4 Jun 2004

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Robin Bowes]

On Fri, June 4, 2004 7:29, Neil Brown said:
>   - Further support for partitionable arrays included "--auto=" option
>     and "auto=" config file entry which instructs mdadm to create the necessary
>     device files after allocating an unused array number.

Neil,

I am interpreting this to mean that I can create, for example, a large (e.g. 1TB) RAID5
array and create smaller partitions (/home, /usr, etc) on top of the RAID5 array? Is
this correct?

Where can I read more about this?

Thanks,

R.
-- 
http://robinbowes.com

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Neil Brown]

On Friday June 4, robin@robinbowes.com wrote:
> On Fri, June 4, 2004 7:29, Neil Brown said:
> >   - Further support for partitionable arrays included "--auto=" option
> >     and "auto=" config file entry which instructs mdadm to create the necessary
> >     device files after allocating an unused array number.
> 
> Neil,
> 
> I am interpreting this to mean that I can create, for example, a large (e.g. 1TB) RAID5
> array and create smaller partitions (/home, /usr, etc) on top of the RAID5 array? Is
> this correct?

Yes, this is correct.

> 
> Where can I read more about this?

Uhm.... Good question.  The are brief hints in
   man mdadm
and
   man mdadm.conf
but I guess I should put something in:
   man 4 md

In recent 2.6 kernels you can create partitionable md arrays.
The is no fixed major number.  The major number can be found by
looking for "mdp" in /proc/devices.
There are 64 minor numbers for each array, allowing up to 63
partitions. 

The simplest way to access this functionality is with

    mdadm --create /dev/md/XX --auto=partition --level= ....

to create, or
 
    mdadm --assemble /dev/md/XX --auto=partition .....

to assemble.  Alternately "auto=partition" can be put in mdadm.conf.

If you do this, it will create device files:
 
    /dev/md/XX   the whole array
    /dev/md/XX1  the first partition
    /dev/md/XX2  the second partition
    /dev/md/XX3  the third partition
    /dev/md/XX4  the fourth partition

More (or less) partitions can be created using e.g. --auto=partition8
to create 8 partition device files.

The arrays will appear in /proc/mdstat as

    md_dN

where N is the number that was chosen.

Then
   cfdisk /dev/md/XX

will allow you to partition the array.

I hope this helps.

NeilBrown



> 
> Thanks,
> 
> R.
> -- 
> http://robinbowes.com

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[me]

> The simplest way to access this functionality is with
> 
>     mdadm --create /dev/md/XX --auto=partition --level= ....
> 

How does in know how big to build the partitions?
How does it know how many partitions?

Can you do it with out building the partitions automatically
and instead it will just build /dev/mdXX and then I can use
fdisk, sdisk, cfdisk... to build my partitions?

Thanks
Jay

RE: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Guy]

You must finish reading what Neil said!

< Then
<    cfdisk /dev/md/XX
< will allow you to partition the array.

He has more info you somehow skipped!
Re-read the email from Neil.


-----Original Message-----
From: linux-raid-owner@vger.kernel.org
[mailto:linux-raid-owner@vger.kernel.org] On Behalf Of me@heyjay.com
Sent: Friday, June 04, 2004 9:31 AM
To: Neil Brown; linux-raid@vger.kernel.org
Subject: Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
Linux

> The simplest way to access this functionality is with
> 
>     mdadm --create /dev/md/XX --auto=partition --level= ....
> 

How does in know how big to build the partitions?
How does it know how many partitions?

Can you do it with out building the partitions automatically
and instead it will just build /dev/mdXX and then I can use
fdisk, sdisk, cfdisk... to build my partitions?

Thanks
Jay
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[me]

I did finish, but the example seemed to show that the --auto would build the
partitions.  I know at the end of the message Neil said then you use cfdisk
but it seemed unnecessary at that point

Jay
----- Original Message ----- 
From: "Guy" <bugzilla@watkins-home.com>
To: <me@heyjay.com>; <linux-raid@vger.kernel.org>
Sent: Friday, June 04, 2004 9:21 AM
Subject: RE: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
Linux


> You must finish reading what Neil said!
>
> < Then
> <    cfdisk /dev/md/XX
> < will allow you to partition the array.
>
> He has more info you somehow skipped!
> Re-read the email from Neil.
>
>
> -----Original Message-----
> From: linux-raid-owner@vger.kernel.org
> [mailto:linux-raid-owner@vger.kernel.org] On Behalf Of me@heyjay.com
> Sent: Friday, June 04, 2004 9:31 AM
> To: Neil Brown; linux-raid@vger.kernel.org
> Subject: Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
> Linux
>
> > The simplest way to access this functionality is with
> >
> >     mdadm --create /dev/md/XX --auto=partition --level= ....
> >
>
> How does in know how big to build the partitions?
> How does it know how many partitions?
>
> Can you do it with out building the partitions automatically
> and instead it will just build /dev/mdXX and then I can use
> fdisk, sdisk, cfdisk... to build my partitions?
>
> Thanks
> Jay
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Neil Brown]

On Friday June 4, me@heyjay.com wrote:
> I did finish, but the example seemed to show that the --auto would build the
> partitions.  I know at the end of the message Neil said then you use cfdisk
> but it seemed unnecessary at that point

Hmmm.... 
I said:

: If you do this, it will create device files:
:  
:     /dev/md/XX   the whole array
:     /dev/md/XX1  the first partition
:     /dev/md/XX2  the second partition
:     /dev/md/XX3  the third partition
:     /dev/md/XX4  the fourth partition

Meaning that it will create "device files" for the partitions, which
is what I said.
However I then said:

: More (or less) partitions can be created using e.g. --auto=partition8
: to create 8 partition device files.

Which first says that "partitions" can be created, and then that
"partition device files" will be created.  

So I clearly wasn't being precise as would be good, though I was
trying :-)

mdadm created device files (just like mknod does).  It does not create
partitions - that is a job for *fdisk.

NeilBrown

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[me]

Sorry for the delayed response.

I think I'm confused about device files and the relationship between device
files and partitions.  I'm going to have to go read some now...

Thanks
Jay
----- Original Message ----- 
From: "Neil Brown" <neilb@cse.unsw.edu.au>
To: <me@heyjay.com>
Cc: "Guy" <bugzilla@watkins-home.com>; <linux-raid@vger.kernel.org>
Sent: Saturday, June 05, 2004 1:08 AM
Subject: Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
Linux


> On Friday June 4, me@heyjay.com wrote:
> > I did finish, but the example seemed to show that the --auto would build
the
> > partitions.  I know at the end of the message Neil said then you use
cfdisk
> > but it seemed unnecessary at that point
>
> Hmmm....
> I said:
>
> : If you do this, it will create device files:
> :
> :     /dev/md/XX   the whole array
> :     /dev/md/XX1  the first partition
> :     /dev/md/XX2  the second partition
> :     /dev/md/XX3  the third partition
> :     /dev/md/XX4  the fourth partition
>
> Meaning that it will create "device files" for the partitions, which
> is what I said.
> However I then said:
>
> : More (or less) partitions can be created using e.g. --auto=partition8
> : to create 8 partition device files.
>
> Which first says that "partitions" can be created, and then that
> "partition device files" will be created.
>
> So I clearly wasn't being precise as would be good, though I was
> trying :-)
>
> mdadm created device files (just like mknod does).  It does not create
> partitions - that is a job for *fdisk.
>
> NeilBrown
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[David Greaves]

I can see why you said that.

my reading is that --auto *creates the major/minor numbers* by which the 
kernel recognises partitions.
It doesn't create the partitions.

I don't have an hdg, but if I did its partitions would be addressed like so:
david@ash:~$ ll /dev/hdg3
brw-rw----    1 root     disk      34,   3 Apr 15  2001 /dev/hdg3

I'd still need cfdisk et al to set the start/end blocks for the device.

Maybe:

--auto=partition creates the device files for partitions on the array 
which may then be manipulated as normal by cfdisk et al

would be a good --help entry?

David

me@heyjay.com wrote:

>I did finish, but the example seemed to show that the --auto would build the
>partitions.  I know at the end of the message Neil said then you use cfdisk
>but it seemed unnecessary at that point
>
>Jay
>----- Original Message ----- 
>From: "Guy" <bugzilla@watkins-home.com>
>To: <me@heyjay.com>; <linux-raid@vger.kernel.org>
>Sent: Friday, June 04, 2004 9:21 AM
>Subject: RE: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
>Linux
>
>
>  
>
>>You must finish reading what Neil said!
>>
>>< Then
>><    cfdisk /dev/md/XX
>>< will allow you to partition the array.
>>
>>He has more info you somehow skipped!
>>Re-read the email from Neil.
>>
>>
>>-----Original Message-----
>>From: linux-raid-owner@vger.kernel.org
>>[mailto:linux-raid-owner@vger.kernel.org] On Behalf Of me@heyjay.com
>>Sent: Friday, June 04, 2004 9:31 AM
>>To: Neil Brown; linux-raid@vger.kernel.org
>>Subject: Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
>>Linux
>>
>>    
>>
>>>The simplest way to access this functionality is with
>>>
>>>    mdadm --create /dev/md/XX --auto=partition --level= ....
>>>
>>>      
>>>
>>How does in know how big to build the partitions?
>>How does it know how many partitions?
>>
>>Can you do it with out building the partitions automatically
>>and instead it will just build /dev/mdXX and then I can use
>>fdisk, sdisk, cfdisk... to build my partitions?
>>
>>Thanks
>>Jay
>>-
>>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
>>the body of a message to majordomo@vger.kernel.org
>>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>
>>-
>>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
>>the body of a message to majordomo@vger.kernel.org
>>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>
>>    
>>
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>  
>

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[me]

Thanks David,

I think you and Neil are saying the same thing. I'm going to do some reading

Jay
----- Original Message ----- 
From: "David Greaves" <david@dgreaves.com>
To: <me@heyjay.com>
Cc: "Guy" <bugzilla@watkins-home.com>; <linux-raid@vger.kernel.org>
Sent: Saturday, June 05, 2004 1:36 AM
Subject: Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
Linux


> I can see why you said that.
>
> my reading is that --auto *creates the major/minor numbers* by which the
> kernel recognises partitions.
> It doesn't create the partitions.
>
> I don't have an hdg, but if I did its partitions would be addressed like
so:
> david@ash:~$ ll /dev/hdg3
> brw-rw----    1 root     disk      34,   3 Apr 15  2001 /dev/hdg3
>
> I'd still need cfdisk et al to set the start/end blocks for the device.
>
> Maybe:
>
> --auto=partition creates the device files for partitions on the array
> which may then be manipulated as normal by cfdisk et al
>
> would be a good --help entry?
>
> David
>
> me@heyjay.com wrote:
>
> >I did finish, but the example seemed to show that the --auto would build
the
> >partitions.  I know at the end of the message Neil said then you use
cfdisk
> >but it seemed unnecessary at that point
> >
> >Jay
> >----- Original Message ----- 
> >From: "Guy" <bugzilla@watkins-home.com>
> >To: <me@heyjay.com>; <linux-raid@vger.kernel.org>
> >Sent: Friday, June 04, 2004 9:21 AM
> >Subject: RE: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
> >Linux
> >
> >
> >
> >
> >>You must finish reading what Neil said!
> >>
> >>< Then
> >><    cfdisk /dev/md/XX
> >>< will allow you to partition the array.
> >>
> >>He has more info you somehow skipped!
> >>Re-read the email from Neil.
> >>
> >>
> >>-----Original Message-----
> >>From: linux-raid-owner@vger.kernel.org
> >>[mailto:linux-raid-owner@vger.kernel.org] On Behalf Of me@heyjay.com
> >>Sent: Friday, June 04, 2004 9:31 AM
> >>To: Neil Brown; linux-raid@vger.kernel.org
> >>Subject: Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under
> >>Linux
> >>
> >>
> >>
> >>>The simplest way to access this functionality is with
> >>>
> >>>    mdadm --create /dev/md/XX --auto=partition --level= ....
> >>>
> >>>
> >>>
> >>How does in know how big to build the partitions?
> >>How does it know how many partitions?
> >>
> >>Can you do it with out building the partitions automatically
> >>and instead it will just build /dev/mdXX and then I can use
> >>fdisk, sdisk, cfdisk... to build my partitions?
> >>
> >>Thanks
> >>Jay
> >>-
> >>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> >>the body of a message to majordomo@vger.kernel.org
> >>More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>
> >>
> >>-
> >>To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> >>the body of a message to majordomo@vger.kernel.org
> >>More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>
> >>
> >>
> >>
> >
> >-
> >To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> >the body of a message to majordomo@vger.kernel.org
> >More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >
> >
> >
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Robin Bowes]

On Fri, June 4, 2004 13:13, Neil Brown said:
> On Friday June 4, robin@robinbowes.com wrote:
>> Where can I read more about this?
>
> Uhm.... Good question.  The are brief hints in
>    man mdadm
> and
>    man mdadm.conf
> but I guess I should put something in:
>    man 4 md

Ok, I've d/l and built the man pages for mdadm. I can't access my linux box at the
moment so I've done this on my WinXP laptop under cygwin. mdadm wouldn't build under
cygwin so I just built and installed the manpages. Here's a patch for the Makefile to
allow the manpages to be built/installed separately with "make man" and "make
install.man":

*** Makefile    Fri Jun  4 07:18:28 2004
--- Makefile.man        Fri Jun  4 15:11:27 2004
***************
*** 60,63 ****
--- 60,64 ----

  all : mdadm mdadm.man md.man mdadm.conf.man
+ man : mdadm.man md.man mdadm.conf.man

  everything: all mdadm.static mdadm.tcc mdadm.uclibc
***************
*** 106,109 ****
--- 107,115 ----
        $(INSTALL) -D -m 644 mdadm.conf.5 $(DESTDIR)$(MAN5DIR)/mdadm.conf.5

+ install.man : mdadm.8 md.4 mdadm.conf.5
+       $(INSTALL) -D -m 644 mdadm.8 $(DESTDIR)$(MAN8DIR)/mdadm.8
+       $(INSTALL) -D -m 644 md.4 $(DESTDIR)$(MAN4DIR)/md.4
+       $(INSTALL) -D -m 644 mdadm.conf.5 $(DESTDIR)$(MAN5DIR)/mdadm.conf.5
+
  clean :
        rm -f mdadm $(OBJS) core *.man mdadm.tcc mdadm.uclibc mdadm.static *.orig
*.porig *.rej *.alt

Also I noticed a typo in the mdadm man page:

CREATE MODE
       Usage: mdadm --create device --chunk=X --level=Y
                   --raid-devices=Z devices

       This  usage will initialise a new md array, associate some devices with
       it, and activate the array.

       This the --auto option is given (as described in  more  detail  in  the ...

The second "This" should be "If" (I'm soooo anal sometimes!)

Anyway, back to the subject in hand...

I have a linux box into which I shall be installed 6 x 250GB SATA drives on a couple of
Promise controllers (3 per controller). I guess these will be named
/dev/sd{a,b,c,d,e,f}.

I then intend to create a couple of mirrored partitions, one to boot from and one for
swap, plus a large RAID5 array with several partitions for /usr,/home, /data, etc.

So, would I do something like:

# cfdisk /dev/sd{a,b,c,d,e,f}
(On each disk create one small partition for the RAID1 array + swap, plus one large
partition for the RAID5)

# grep mdp < /proc/devices
(Get the major numbers of the md devices - will I have more than one, eg XX1, XX2?)

# mdadm --create /dev/md/XX1 --auto=partition2 --level=1 --raid-devices=6 /dev/sd[abcdef]1
(create a RAID1 array from all six drives containing two partitions )

# mdadm --create /dev/md/XX2 --auto=partition6 --level=5 --raid-devices=6 /dev/sd[abcdef]2
(create a RAID5 array from all six drives containing 6 partitions)

# cfdisk /dev/md/XX1
(create /boot and swap partitions)

# cfdisk /dev/md/XX2
(create partitions on RAID5 array)

One issue:

 - Would it be better to create the mirrored arrays as separate, "standard" md arrays of
three drives (2 plus spare). Would I do this by 1) creating the necessary partitions,
2) building the array like this:

# mdadm --create /dev/md0 --level=1 --raid-devices=2 /dev/sd[ad]1 --spare-devices=1
/dev/sdb1

# mdadm --create /dev/md1 --level=1 --raid-devices=2 /dev/sd[cf]1 --spare-devices=1
/dev/sde1

Am I somewhere near the mark?

Cheers,

R.
-- 
http://robinbowes.com

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[David Greaves]

Robin Bowes wrote:

>Anyway, back to the subject in hand...
>
>I have a linux box into which I shall be installed 6 x 250GB SATA drives on a couple of
>Promise controllers (3 per controller). I guess these will be named
>/dev/sd{a,b,c,d,e,f}.
>  
>
<grin>
I just installed a single controller (TX4) and 3 250Gb drives...
the 4th, as I type, is with a courier, lost somewhere in Berkshire...

>I then intend to create a couple of mirrored partitions, one to boot from and one for
>swap, plus a large RAID5 array with several partitions for /usr,/home, /data, etc.
>
>So, would I do something like:
>  
>
<snip>
yes.

But any reason not to use lvm2? (anyone?)

If you fancy a cut'n'paste of my commands and output, look here: 
http://www.mythtv.info/moin.cgi/AdministrationSoftware_2fLvmRaid

Also talks you through setting up lvm2 on your new raid 5 array.
Later, when you run out of space (!) you can add another array and or 
shuffle partitions about using lvm.

David

Re: ANNOUNCE: mdadm 1.6.0 - A tool for managing Soft RAID under Linux

[Robin Bowes]

On Fri, June 4, 2004 16:22, David Greaves said:
>
> But any reason not to use lvm2? (anyone?)

TO abstract your suggestion a little, I'm considering using EVMS to manage my storage.

I believe EVMS does not support lvm2 yet.

> If you fancy a cut'n'paste of my commands and output, look here:
> http://www.mythtv.info/moin.cgi/AdministrationSoftware_2fLvmRaid

Thanks. Useful stuff.

R.
-- 
http://robinbowes.com