LinuxQuestions.org - Demystifying /dev/mapper

- Linux - Newbie (https://www.linuxquestions.org/questions/linux-newbie-8/)

- - Demystifying /dev/mapper (https://www.linuxquestions.org/questions/linux-newbie-8/demystifying-dev-mapper-4175619724/)

Demystifying /dev/mapper

I mounted /home/michael/devicefilexxx at /home/michael/mountpointxxx as shown below. Then I looked at /etc/mtab to see what it looked like before manually editing /etc/fstab. /dev/mapper/centos-root? What is that? Did a bit of research, and found that I am evidently using Logical Volume Management. Okay, good I guess, but have a couple of questions:

How can I have device /dev/mapper/centos-root mounted both at / and /home/michael/mountpointxxx?
How can I identify specifically what is mounted at /home/michael/mountpointxxx?
How do I permanently mount this device? Surely, I cannot simply add /dev/mapper/centos-root /home/michael/mountpointxxx xfs rw,relatime,attr2,inode64,noquota 0 0 to /etc/fdisk as /dev/mapper/centos-root will not uniquely identify the device.
Should I be using a special tool to permanently mount devices instead of just editing /etc/fstab?
What is with /dev/dm-0, /dev/dm-1, and /dev/mapper/control?
Please explain the output of lsblk and how it pertains to the other outputs.
Based on my above questions (or more likely, lack of some pointed question) and displayed command outputs below, what else should I be asking and what other commands should I be running to view my system configuration?

Thank you

Code:

[michael@vps ~]$ mkdir /home/michael/devicefilexxx

[michael@vps ~]$ mkdir /home/michael/mountpointxxx

[michael@vps ~]$ sudo mount --bind /home/michael/devicefilexxx /home/michael/mountpointxxx



[michael@vps ~]$ cat /etc/mtab | grep xxx

/dev/mapper/centos-root /home/michael/mountpointxxx xfs rw,relatime,attr2,inode64,noquota 0 0



[michael@vps ~]$ mount

sysfs on /sys type sysfs (rw,nosuid,nodev,noexec,relatime)

proc on /proc type proc (rw,nosuid,nodev,noexec,relatime)

devtmpfs on /dev type devtmpfs (rw,nosuid,size=1009596k,nr_inodes=252399,mode=755)

securityfs on /sys/kernel/security type securityfs (rw,nosuid,nodev,noexec,relatime)

tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)

devpts on /dev/pts type devpts (rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000)

tmpfs on /run type tmpfs (rw,nosuid,nodev,mode=755)

tmpfs on /sys/fs/cgroup type tmpfs (ro,nosuid,nodev,noexec,mode=755)

cgroup on /sys/fs/cgroup/systemd type cgroup (rw,nosuid,nodev,noexec,relatime,xattr,release_agent=/usr/lib/systemd/systemd-cgroups-agent,name=systemd)

pstore on /sys/fs/pstore type pstore (rw,nosuid,nodev,noexec,relatime)

cgroup on /sys/fs/cgroup/pids type cgroup (rw,nosuid,nodev,noexec,relatime,pids)

cgroup on /sys/fs/cgroup/net_cls,net_prio type cgroup (rw,nosuid,nodev,noexec,relatime,net_prio,net_cls)

cgroup on /sys/fs/cgroup/cpu,cpuacct type cgroup (rw,nosuid,nodev,noexec,relatime,cpuacct,cpu)

cgroup on /sys/fs/cgroup/perf_event type cgroup (rw,nosuid,nodev,noexec,relatime,perf_event)

cgroup on /sys/fs/cgroup/devices type cgroup (rw,nosuid,nodev,noexec,relatime,devices)

cgroup on /sys/fs/cgroup/blkio type cgroup (rw,nosuid,nodev,noexec,relatime,blkio)

cgroup on /sys/fs/cgroup/cpuset type cgroup (rw,nosuid,nodev,noexec,relatime,cpuset)

cgroup on /sys/fs/cgroup/memory type cgroup (rw,nosuid,nodev,noexec,relatime,memory)

cgroup on /sys/fs/cgroup/freezer type cgroup (rw,nosuid,nodev,noexec,relatime,freezer)

cgroup on /sys/fs/cgroup/hugetlb type cgroup (rw,nosuid,nodev,noexec,relatime,hugetlb)

configfs on /sys/kernel/config type configfs (rw,relatime)

/dev/mapper/centos-root on / type xfs (rw,relatime,attr2,inode64,noquota)

systemd-1 on /proc/sys/fs/binfmt_misc type autofs (rw,relatime,fd=34,pgrp=1,timeout=300,minproto=5,maxproto=5,direct)

mqueue on /dev/mqueue type mqueue (rw,relatime)

hugetlbfs on /dev/hugepages type hugetlbfs (rw,relatime)

debugfs on /sys/kernel/debug type debugfs (rw,relatime)

/dev/sda1 on /boot type xfs (rw,relatime,attr2,inode64,noquota)

tmpfs on /run/user/1000 type tmpfs (rw,nosuid,nodev,relatime,size=204060k,mode=700,uid=1000,gid=1000)

/dev/mapper/centos-root on /home/michael/mountpointxxx type xfs (rw,relatime,attr2,inode64,noquota)



[michael@vps ~]$ cat /etc/fstab



#

# /etc/fstab

# Created by anaconda on Fri Apr  8 14:15:42 2016

#

# Accessible filesystems, by reference, are maintained under '/dev/disk'

# See man pages fstab(5), findfs(8), mount(8) and/or blkid(8) for more info

#

/dev/mapper/centos-root /                      xfs    defaults        1 1

UUID=362355d4-e5da-44de-bf5c-5ce92cf43888 /boot                  xfs    defaults        1 2

/dev/mapper/centos-swap swap                    swap    defaults        0 0



[michael@vps ~]$ lsblk

NAME            MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT

sda              8:0    0  80G  0 disk

├─sda1            8:1    0  500M  0 part /boot

└─sda2            8:2    0 79.5G  0 part

  ├─centos-swap 253:0    0    2G  0 lvm  [SWAP]

  └─centos-root 253:1    0 77.5G  0 lvm  /

sr0              11:0    1 1024M  0 rom



[michael@vps ~]$ ls -l /dev/mapper

total 0

lrwxrwxrwx 1 root root      7 Dec 11 04:09 centos-root -> ../dm-1

lrwxrwxrwx 1 root root      7 Dec 11 04:09 centos-swap -> ../dm-0

crw------- 1 root root 10, 236 Dec 11 04:09 control

[michael@vps ~]$ ls -l /dev/mapper/../dm-1

brw-rw---- 1 root disk 253, 1 Dec 11 04:09 /dev/mapper/../dm-1

[michael@vps ~]$ ls -l /dev/mapper/control

crw------- 1 root root 10, 236 Dec 11 04:09 /dev/mapper/control



[michael@vps ~]$ sudo lvdisplay

  --- Logical volume ---

  LV Path                /dev/centos/swap

  LV Name                swap

  VG Name                centos

  LV UUID                M1qwjs-A6iN-BSMr-CQeA-2VLQ-IuLd-eM1smR

  LV Write Access        read/write

  LV Creation host, time localhost, 2016-04-08 14:15:41 +0000

  LV Status              available

  # open                2

  LV Size                2.00 GiB

  Current LE            512

  Segments              1

  Allocation            inherit

  Read ahead sectors    auto

  - currently set to    8192

  Block device          253:0



  --- Logical volume ---

  LV Path                /dev/centos/root

  LV Name                root

  VG Name                centos

  LV UUID                ejjUQX-f22h-HH82-3Cg8-6Tc5-4qtP-bJt0z3

  LV Write Access        read/write

  LV Creation host, time localhost, 2016-04-08 14:15:42 +0000

  LV Status              available

  # open                1

  LV Size                77.51 GiB

  Current LE            19842

  Segments              1

  Allocation            inherit

  Read ahead sectors    auto

  - currently set to    8192

  Block device          253:1



[michael@vps ~]$ sudo vgdisplay

  --- Volume group ---

  VG Name              centos

  System ID

  Format                lvm2

  Metadata Areas        1

  Metadata Sequence No  6

  VG Access            read/write

  VG Status            resizable

  MAX LV                0

  Cur LV                2

  Open LV              2

  Max PV                0

  Cur PV                1

  Act PV                1

  VG Size              79.51 GiB

  PE Size              4.00 MiB

  Total PE              20354

  Alloc PE / Size      20354 / 79.51 GiB

  Free  PE / Size      0 / 0

  VG UUID              zPAZLs-6QQe-Tnvz-ijly-mKQu-MFk0-ie0kho



[michael@vps ~]$ sudo pvdisplay

  --- Physical volume ---

  PV Name              /dev/sda2

  VG Name              centos

  PV Size              79.51 GiB / not usable 2.00 MiB

  Allocatable          yes (but full)

  PE Size              4.00 MiB

  Total PE              20354

  Free PE              0

  Allocated PE          20354

  PV UUID              AVJtne-FSKj-8Guj-kv7U-Jis1-x5Nk-Sp60zp



[michael@vps ~]$

You can't (in any sensible sense). :)

LVM works by taking a bunch of Physical Volumes (PF).
You separate the bunch of PF's into one or more Volume Groups (VG).
Each VG is divided into several Logical Files (LF).
Each LF is equivalent to a normal hard disk partition, except that it is easily resizeable.

Note that a disk error, anywhere, can effectively destroy your entire disk setup.

But it is a good system if you would otherwise be frequently resizing partitions, or needing to have partitions which are bigger than any existing disk.
(I use it for squeezing backups onto a single large disk).

Have a read of p12 of https://fullcirclemagazine.org/issue-80/

Thanks Jeremy, I can't what? Hopefully not mount /home/michael/devicefilexxx at /home/michael/mountpointxxx and have it remain mounted after a reboot. Doing so is my immediate goal, and hoping to benefit from a bit of education in the process.

It wouldn't make sense for Q1. :)

/dev/mapper/centos-root could uniquely identify / (in the same way that /dev/sda1 could uniquely identify /).

There is no reason why you cannot mount /dev/mapper/centos-home to a mountpoint such as /mnt/home since you can now have a giant /home directory which spans multiple disks.
Obviously that's not obligatory though.

BTW I suspect the system doesn't handle hyphens well - I suggest you use underscores.

I don't have access to a LVM set-up at the moment...

If you want to play with LVM, you can do so with a USB key of a reasonable size.
In Debian, I could get all the LVM commands by entering (as root)

Code:

lvm

help

For further help I would enter a command followed by -h

Thanks again Jeremy,

I am glad Q1 didn't make sense to you because it made no sense to me! :)

Hopefully, the hyphens are not an issue because the server in question is being hosted by 1&1 and they set it up.

I can (and have mounted the device as shown at the very beginning of my original post), but don't know how to make it permanently remain after a reboot. :(

The hyphen is not important. When I used it:-
If you look at names outside of lvm all hyphens get repeated,
but when you issue commands inside lvm you only use the single hyphen.
Actually, I think I now sort of understand it???

You actually chose a LV Path of /dev/centos/root
However, your fstab file calls this /dev/mapper/centos-root
In some horrible way, I would guess that the mapper gets removed and the hyphen is transliterated into a /.
This makes lvm happy.

But don't expect too much from non-lvm commands like lsblk - dm-0 etc just correspond to bits of storage.
If you want to see what's mounted,

Code:

mount

Code:

df -h

seem to be useful.

Thanks again Jeremy, But how can I make the mount remain after the machine is rebooted? Typically, one would edit /etc/fstab, no? And put what as the device? /dev/mapper/centos-root and somehow it knows what I was thinking? While that would be impressive, I don't think Linux is that smart. Somehow I need to specify both the device path and mount path...

Have to admit that I don't rely on fstab for my backups.
I use a script to mount and unmount LV's as they are needed.

But that doesn't mean fstab wouldn't work - especially if you used UUID's to identify the LV's.

To add a bind mount entry in fstab, just follow manual
man mount

Code:

  The bind mounts.

          Since Linux 2.4.0 it is possible to remount  part  of  the  file

          hierarchy somewhere else. The call is

                mount --bind olddir newdir

          or shortoption

                mount -B olddir newdir

          or fstab entry is:

                /olddir /newdir none bind

Quote:

How can I have device /dev/mapper/centos-root mounted both at / and /home/michael/mountpointxxx?

A bind mount is possible.
The bind driver redirects all access to the primary mount.

Quote:

How can I identify specifically what is mounted at ?

df /home/michael/mountpointxxx shows the device.
You can run df on any(!) file, it gives you the filesystem (mount point, device, size, usage)

Quote:

How do I permanently mount this device? Surely, I cannot simply add /dev/mapper/centos-root /home/michael/mountpointxxx xfs rw,relatime,attr2,inode64,noquota 0 0 to /etc/fdisk as /dev/mapper/centos-root will not uniquely identify the device.

Why not? lvs or lvdisplay show the device. I suggest to do the fstab entry first, then run mount /home/michael/mountpointxxx that looks up the other things from the fstab.

Quote:

Should I be using a special tool to permanently mount devices instead of just editing /etc/fstab?

No. A simple text editor like vi is good.

Quote:

What is with /dev/dm-0, /dev/dm-1, and /dev/mapper/control?
Please explain the output of lsblk and how it pertains to the other outputs.

lsblk and pvs or pvdisplay display the physical devices (disks and disk slices).
The next stage is the volume group, displayed with vgs or vgdisplay
The last step is the logical volume, displayed with lvs or lvdisplay

Redhat has very decent documentation about lvm2.

Quote:

--- Logical volume ---
LV Path /dev/centos/root
LV Name root
VG Name centos

This already explains the mess.

in comparison

Quote:

ASUS-G75VW /home/roman # lvs
LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert
lv_root_volume_SanDisk vg_root_volume_SanDisk -wi-ao---- 111,63g

It is good practise to name the stuff so a human can easily read it.
Naming stuff like centos, root, swap is just bad working practise. Regardless if an installer did it or a human.

I have quite long names, but I can easily say what it is

lv_root_volume_SanDisk => logical volume = lv; root_volume ... /; SanDisk ... manufacturer of the drive who I can blame
vg_root_volume_SanDisk => volume group = vg; root_volume ... /; SanDisk ... manufacturer of the drive who I can blame

--

Quote:

ASUS-G75VW /home/roman # ls -alh /dev/mapper/
total 0
drwxr-xr-x 2 root root 100 Dez 19 13:12 .
drwxr-xr-x 21 root root 3,7K Dez 19 13:12 ..
crw------- 1 root root 10, 236 Dez 19 14:12 control
lrwxrwxrwx 1 root root 7 Dez 19 13:12 root_vg_root_volume_SanDisk-lv_root_volume_SanDisk-root -> ../dm-1
lrwxrwxrwx 1 root root 7 Dez 19 13:12 vg_root_volume_SanDisk-lv_root_volume_SanDisk -> ../dm-0

Let's be very unprecise.

vg_root_volume_SanDisk-lv_root_volume_SanDisk => as you can see they are put together with "-" => vg + "-" + lv

I assume my old initramfs and gentoo eudev + openrc recreates it again => lrwxrwxrwx 1 root root 7 Dez 19 13:12 root_vg_root_volume_SanDisk-lv_root_volume_SanDisk-root -> ../dm-1 => i do see a boot error but I do not care. I do backups regularly and recreate the file system during backup stage. less fragmentation, no need for file system checks. => lazy dirty approach. <= i do not care for the init cry about not beeing able to do a file system check.

--

as for the dm stuff => https://wiki.gentoo.org/wiki/Device-mapper

I only use lvm2 + luks. For several reasons

--

luckily lvm2 provides rename tools. You may use em. CAreful to fix your boot process.

--

Quote:

What is with /dev/dm-0, /dev/dm-1, and /dev/mapper/control?

Let's take my example. Those are links. => man ln => ln - make links between files (out of scope ... there are decent guides on soft and hard links)

--

https://unix.stackexchange.com/quest...apper-in-fdisk

https://en.wikipedia.org/wiki/Device_mapper

Quote:

Applications (like LVM2 and EVMS) that need to create new mapped devices talk to the device mapper via the libdevmapper.so shared library, which in turn issues ioctls to the /dev/mapper/control device node.[4]

--

http://"https://askubuntu.com/questi...-and-etc-mtab"

I assume some sort of automounter in your distro screwed up.
I did not used mtab for quite a while.

Judging from your mount output, it seems you have your "/" mounted two times in your system, once as /, once as subfolder in your home directory

Quote:

You actually chose a LV Path of /dev/centos/root
However, your fstab file calls this /dev/mapper/centos-root
In some horrible way, I would guess that the mapper gets removed and the hyphen is transliterated into a /.
This makes lvm happy.

What the ? The two forms /dev/vg_name/lv_name and /dev/mapper/vg_name-lv_name both point to the same /dev/dm node. They are alternate forms. There is nothing "horrible" about it.

Quote:

LVM works by taking a bunch of Physical Volumes (PF).
You separate the bunch of PF's into one or more Volume Groups (VG).
Each VG is divided into several Logical Files (LF).
Each LF is equivalent to a normal hard disk partition, except that it is easily resizeable.

Physical Volumes are PV (not PF).
Volume Groups are divided into Logical Volumes (LV).

Quote:

But how can I make the mount remain after the machine is rebooted? Typically, one would edit /etc/fstab, no? And put what as the device? /dev/mapper/centos-root and somehow it knows what I was thinking? While that would be impressive, I don't think Linux is that smart. Somehow I need to specify both the device path and mount path...

Again what the ? You would just use /dev/mapper/centos-root in fstab as that is the link to the /dev/dm-n node. Simple as that. Users use links with the VG and LV names in them. These links with the same names are created at every boot.

Quote:

Surely you can. /dev/mapper/centos-root will uniquely identify the device as there is only one LVM VG named centos and it has only one LV named root. Therefore the link named /dev/mapper/centos-root will be unique. This link will point to the correct /dev/dm device at every boot.

Thanks tofino surfer. Been a while since I've been to Tofino. Surf good?

Surly I am missing something. There are two unique paths /home/michael/mountpoint1 and /home/michael/mountpoint2. Neither of them point to / which I expected /dev/mapper/centos-root to be and both paths are not equal so can't point to the same place. Maybe I am confusing myself with the use of the word "point"? I agree /dev/mapper/centos-root uniquely locates a place where they might be located within, but not specifically where they are located within that space.

Code:

[michael@vps ~]$ mkdir /home/michael/devicefile1

[michael@vps ~]$ mkdir /home/michael/devicefile2

[michael@vps ~]$ mkdir /home/michael/mountpoint1

[michael@vps ~]$ mkdir /home/michael/mountpoint2

[michael@vps ~]$ sudo mount --bind /home/michael/devicefile1 /home/michael/mountpoint1

[michael@vps ~]$ sudo mount --bind /home/michael/devicefile2 /home/michael/mountpoint2

[michael@vps ~]$ cat /etc/mtab | grep mountpoint

/dev/mapper/centos-root /home/michael/mountpoint1 xfs rw,relatime,attr2,inode64,noquota 0 0

/dev/mapper/centos-root /home/michael/mountpoint2 xfs rw,relatime,attr2,inode64,noquota 0 0

[michael@vps ~]$

Quote:

Originally Posted by tofino_surfer (Post 5795839)

There is nothing "horrible" about it.

Any guide says use something human readable. Which makes things clear.

There are certain working principles in life. Go on and name it as short as possible so no one has a clue.
Sarcasm: Really good working habit naming it centos, swap, ... lol

When you read the first post you can see the poster has no glue about it.
It is more obvious and clear when you do the thing I suggested.
I read the hole redhat docs about lvm2 in the past.
I read other posts, some were smart some were not.
Than I adapted my box so it would boot. even my initramfs.

I prefer human readable names instead of using pen and paper to write down what is what before i act.
I use pen and paper because sda2 sdb4 are something which i could easily mix up. and when you mix up, you may disk destroyer (= dd) your hole data for example. And there is still the risk on next reboot the kernel mix up the device tree, as in the past

--

I found something to add to this topic

https://forums.gentoo.org/viewtopic-...ighlight-.html

Quote:

well, you can kind of get away not putting the / filesystem in fstab. After all - in order to read the fstab, / must already be mounted.

you are probably passing that info via your bootloader as kernel parameter, and/or through initramfs, which mounts / for you.

of course, you should still put it in fstab in the end, if only to check that the correct thing is mounted and with the options.

This explains my double entry.

This forums.gentoo.org post also explains why the volume group was auto mounted in his home folder. Supposely missing fstab entry or wrong fstab entry. so the automounter mounted it in /home.

Quote:

Originally Posted by tofino_surfer (Post 5795839)

False.

First of all you need something to boot your computer:

For example: The usual layout with sda2 as suggested in the gentoo handbook with vfat.
Than you need an initramfs and an adapted kernel as linux can not boot on its own from lvm2, with or without luks.
(there are claims that it can boot from lvm2, luks whatever, grub or the kernel, but i never saw a working example, just claims it can, so it can't until someone shows me, it does itself, without a bootloader, without an initramfs! it has a reason why the gentoo handbook still sticks to that layout!)
Than you need the rootfs, realroot parameter which was former called root parameter.

As https://forums.gentoo.org/viewtopic-...ighlight-.html explained and I also know from experience the / entry in fstab is not needed anymore. At this point of stage root is already mounted and fstab gets parsed later

I recommend bothering in how a box boots. I did adapt, rewrite (call it what you want it,these days it is called code reuse) my initramfs. I built my own kernels. I recommend step away from the newbie distros and dive deeper.

--

Go on criticise me. But criticising me for a decent working habit and proper naming things?