Re: LVM support

["Tomáš Čech" <sleep_walker@gnu.org>]

[-- Attachment #1: Type: text/plain, Size: 7229 bytes --]

On Tue, Apr 21, 2015 at 05:52:33PM +0200, Ludovic Courtès wrote:
>Tomáš Čech <sleep_walker@gnu.org> skribis:
>
>> On Thu, Apr 16, 2015 at 02:47:52PM +0200, Ludovic Courtès wrote:
>>>Tomáš Čech <sleep_walker@gnu.org> skribis:
>>>
>>>> On Wed, Apr 15, 2015 at 02:32:14PM +0200, Ludovic Courtès wrote:
>>>
>>>[...]
>>>
>>>>>Sorry I’m not really familiar with LVM.
>>>>
>>>> It's implemented using device mapper but instead of mapping one block
>>>> device to another you map one block device to whole group (like
>>>> playground where you can do anything).
>>>
>>>What do you mean by “whole group”?  A tree under /dev/mapper?
>>
>> From device node POV it generates
>> /dev/<volume_group_name>/<logical_volume_name> and it also creates
>> /dev/mapper/<volume_group_name>-<logical_volume_name> and
>> /dev/dm-<number>.
>
>OK.
>
>> From block device perspective it adds another level of "partitioning"
>> to "physical volume" partitions. You gather block devices (can be
>> partitions, disks, anything), create volume group to join the space
>> into one entity and then create logical volumes without caring where
>> it really is. Logical volumes are useful for resizing, adding and
>> removing filesystems - it has always the same device node.
>
>Yes, that part I knew.  ;-)
>
>
>[...]
>
>> --- a/gnu/system.scm
>> +++ b/gnu/system.scm
>> @@ -41,6 +41,7 @@
>>    #:use-module (gnu packages compression)
>>    #:use-module (gnu packages firmware)
>>    #:autoload   (gnu packages cryptsetup) (cryptsetup)
>> +  #:autoload   (gnu packages linux) (lvm2/static)
>>    #:use-module (gnu services)
>>    #:use-module (gnu services dmd)
>>    #:use-module (gnu services base)
>> @@ -86,7 +87,9 @@
>>              %base-packages
>>              %base-firmware
>>
>> -            luks-device-mapping))
>> +            luks-device-mapping
>> +            lvm-mapping
>> +            lvm-mapping-used?))
>>
>>  ;;; Commentary:
>>  ;;;
>> @@ -208,6 +211,27 @@ file."
>>     (open open-luks-device)
>>     (close close-luks-device)))
>>
>> +(define (logical-volume-group-activate source target)
>> +  #~(zero? (system* (string-append #$lvm2/static "/sbin/lvm.static")
>> +                         "vgchange" "--activate" "y" #$target)))
>> +
>> +(define (logical-volume-group-deactivate source target)
>> +  #~(zero? (system* (string-append #$lvm2/static "/sbin/lvm.static")
>> +                    "vgchange" "--activate" "n" #$target)))
>> +
>> +(define (lvm-mapping-used? devices)
>> +  (not
>> +   (null? (filter
>> +           (lambda (md)
>> +             (eq? (mapped-device-type md)
>> +                  lvm-mapping))
>> +           devices))))
>> +
>> +(define lvm-mapping
>> +  (mapped-device-kind
>> +   (open  logical-volume-group-activate)
>> +   (close logical-volume-group-deactivate)))
>
>This looks good to me!
>
>So I would declare
>
>  (mapped-device
>    (source "/dev/sda")
>    (target "volume_group_name-logical_volume_name")
>    (kind lvm-device-mapping))
>
>and that would give me
>/dev/mapper/volume_group_name-logical_volume_name, right?

Volume group can be on multiple block devices. For now I rely on autodetect
abilities of LVM.

So you would declare:

(mapped-device
  (source "") ; irrelevant for LVM
  (target "volume_group_name")
  (type lvm-mapping))

and that would give you
/dev/mapper/volume_group_name-some_volume
/dev/mapper/volume_group_name-other_volume
...

and more conveniently
/dev/volume_group_name/some_volume
/dev/volume_group_name/other_volume
...

>
>>  (define (other-file-system-services os)
>>    "Return file system services for the file systems of OS that are not marked
>>  as 'needed-for-boot'."
>> @@ -267,7 +291,10 @@ from the initrd."
>>          (file-systems (operating-system-file-systems os)))
>>     (filter (lambda (md)
>>               (let ((user (mapped-device-user md file-systems)))
>> -               (and user (file-system-needed-for-boot? user))))
>> +               (or
>> +                (and user (file-system-needed-for-boot? user))
>> +                (and (eq? (mapped-device-type md)
>> +                          lvm-mapping)))))
>>             devices)))
>
>I don’t think it’s necessary: if a ‘file-system’ object has
>"/dev/mapper/volume_group_name-logical_volume_name" has its ‘device’
>field, then this device mapping will automatically be recognized as
>needed-for-boot, won’t it?

Yes, you're right, this chunk shouldn't be needed at all. Good catch!

>> --- a/gnu/system/linux-initrd.scm
>> +++ b/gnu/system/linux-initrd.scm
>> @@ -25,6 +25,7 @@
>>                  #:select (%store-prefix))
>>    #:use-module ((guix derivations)
>>                  #:select (derivation->output-path))
>> +  #:use-module (gnu system)
>>    #:use-module (gnu packages cpio)
>>    #:use-module (gnu packages compression)
>>    #:use-module (gnu packages linux)
>> @@ -212,6 +213,9 @@ loaded at boot time in the order in which they appear."
>>                    file-systems)
>>              (list e2fsck/static)
>>              '())
>> +      ,@(if (lvm-mapping-used? mapped-devices)
>> +            (list lvm2/static)
>> +            '())
>>        ,@(if volatile-root?
>>              (list unionfs-fuse/static)
>>              '())))
>> @@ -241,7 +245,19 @@ loaded at boot time in the order in which they appear."
>>
>>           (boot-system #:mounts '#$(map file-system->spec file-systems)
>>                        #:pre-mount (lambda ()
>> -                                    (and #$@device-mapping-commands))
>> +                                    (and #$@device-mapping-commands
>> +                                         ;; If we activated any volume group, we
>> +                                         ;; need to ensure that device nodes are
>> +                                         ;; created.  Add code here to call it
>> +                                         ;; once for all activations.
>> +                                         #$(when (lvm-mapping-used? mapped-devices)
>> +                                             #~(zero?
>> +                                                (system* (string-append
>> +                                                          #$lvm2/static
>> +                                                          "/sbin/lvm.static")
>> +                                                         "vgscan"
>> +                                                         "--mknodes")))))
>
>So ‘lvm vgchange --activate y’ does not create /dev nodes?

Right.

>Would it be possible to change the command returned by
>‘logical-volume-group-activate’ to somehow create the nodes?  That would
>be ideal.

There are two actions needed to be taken:
1] volume group activation
2] creation of nodes

This design choice does as many 1] as needed and 2] once in the end.

I could do always 1] and 2] for every volume group, but I didn't find it nice,
since previous 2] calls are useless only slowing down the process. Do you
really think I should change it?

Thanks for your review.

S_W

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]