Skip to content


Second impressions on 1and1 (again, or how to deal with LVM)

“mount: unknown filesystem type ‘LVM2_member'”
That’s the error message that eventually allowed me to find the solution.

As I mentioned previously (twice already), when you rent a cloud server at 1and1, and you install Linux on it (well, at least Ubuntu), you obtain a weird partitioning scheme. Such as:

# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/xvda1            3.7G  1.2G  2.6G  31% /
none                  490M  140K  490M   1% /dev
none                  500M  8.0K  500M   1% /dev/shm
none                  500M     0  500M   0% /tmp
none                  500M   60K  500M   1% /var/run
none                  500M     0  500M   0% /var/lock
none                  500M     0  500M   0% /lib/init/rw
/dev/mapper/vg00-usr  4.0G  441M  3.6G  11% /usr
/dev/mapper/vg00-var  4.0G  215M  3.8G   6% /var
/dev/mapper/vg00-home 4.0G  4.3M  4.0G   1% /home

So, in total much less than the 100 GB you get in the smallest case. I tried to follow their (tiny) FAQ on the subject, but seriously this wasn’t worth going farther than:

# fdisk -l

Disk /dev/xvda: 100.0 GB, 100000595968 bytes
255 heads, 63 sectors/track, 12157 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

    Device Boot      Start         End      Blocks   Id  System
/dev/xvda1               1         487     3911796   83  Linux
/dev/xvda2             488         731     1959930   82  Linux swap / Solaris
/dev/xvda3             732       12157    91779345   8e  Linux LVM

(I’ll be trying to provide full output as I guess that’s what’s most helpful if you’re struggling to do just what I wanted to do, i.e. just get the GBs you paid for)
I just tried to mount that big /dev/xvda3 thing, but it failed:

# cd mnt
# ls
# mkdir part1
# mount /dev/xvda3 /mnt/part1
mount: unknown filesystem type 'LVM2_member'

But this error was not in vain, for searching for it brought me the solution: that thing is an LVM volume, and there’s a particular way to deal with it.
First you can (should!) check out your LVM volumes *and groups*:

# pvs
  PV         VG   Fmt  Attr PSize  PFree
  /dev/xvda3 vg00 lvm2 a-   87.52g 75.52g

Here it’s simple, we have one volume, /dev/xvda3, which contains one group, vg00, which has a size of 87.52GiB (among which 75.52GiB unassigned aka free). Looks familiar? You bet, remember earlier:

/dev/mapper/vg00-usr  4.0G  441M  3.6G  11% /usr
/dev/mapper/vg00-var  4.0G  215M  3.8G   6% /var
/dev/mapper/vg00-home 4.0G  4.3M  4.0G   1% /home

Now you’ll want to check out what’s inside that group:

# lvdisplay /dev/vg00
  --- Logical volume ---
  LV Name                /dev/vg00/usr
  VG Name                vg00
  LV UUID                ElBhAu-r2PF-g1sV-UICK-T7Ao-TZmC-uKv0sj
  LV Write Access        read/write
  LV Status              available
  # open                 1
  LV Size                4.00 GiB
  Current LE             1024
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           251:0

  --- Logical volume ---
  LV Name                /dev/vg00/var
  VG Name                vg00
  LV UUID                gz0jhl-7uzG-1wqj-b1Ok-Mqjk-rWlC-ZHBBhy
  LV Write Access        read/write
  LV Status              available
  # open                 1
  LV Size                4.00 GiB
  Current LE             1024
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           251:1

  --- Logical volume ---
  LV Name                /dev/vg00/home
  VG Name                vg00
  LV UUID                2C9KvC-Dcmx-4C6N-T7cg-NzAr-WVOi-KS2xc1
  LV Write Access        read/write
  LV Status              available
  # open                 1
  LV Size                4.00 GiB
  Current LE             1024
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           251:2

It’s like the jackpot, because from all the posts I read about LVM (see sources at the end), it seems that the hard part with those LVM volumes is to mount them. Here (on that 1and1 cloud server) we’re totally in luck: they’re already mounted! We just need to find out how to resize them, and this looks fairly easy: lvextend and lvreduce… Now, before going further, I’d like to warn you about something: extending is most likely totally harmless (and super-fast), reducing is probably more risky, or at least more tedious (you need to unmount the logical volume to do it “safely”, and I’m not sure how safe this is).
So the winning strategy, in my opinion, is maybe to try extending an LVM with 1GiB just to see if it works fine, and then wait until you have a better idea of which volume needs space before extending them again (maybe also wait till that volume gets low on free space – but don’t wait too much, it would be stupid to lose data because of lack of assigned space while there’s so much free ^^). Avoid extending the volumes alternatively and regularly to avoid fragmenting them. Avoid assigning too much too fast to avoid having to reduce a volume later (you’ll see below how scary it sounds when you try it ;)).
So, here are the 2 commands and their outputs:

# lvextend -L+1G /dev/vg00/home
  Extending logical volume home to 5.00 GiB
  Logical volume home successfully resized
# lvreduce -L-1G /dev/vg00/home
  WARNING: Reducing active and open logical volume to 4.00 GiB
  THIS MAY DESTROY YOUR DATA (filesystem etc.)
Do you really want to reduce home? [y/n]: n
  Logical volume home NOT reduced
  Command failed with status code 5.

(yeah, I aborted that one ^^)

Now that you’ve extended the volume, you’ll have to extend the filesystem first. Something you should know before going on if you’re doing this on 1and1 like me: their cloud servers use XFS, and you can’t shrink an XFS file system so once you’ve expanded it there’s no easy way back. Here we go, for instance for /home (the LVM volume needs to be mounted):

# xfs_growfs /home
meta-data=/dev/mapper/vg00-home  isize=256    agcount=8, agsize=131072 blks
         =                       sectsz=512   attr=0
data     =                       bsize=4096   blocks=1048576, imaxpct=25
         =                       sunit=0      swidth=0 blks
naming   =version 2              bsize=4096   ascii-ci=0
log      =internal               bsize=4096   blocks=2560, version=1
         =                       sectsz=512   sunit=0 blks, lazy-count=0
realtime =none                   extsz=65536  blocks=0, rtextents=0
data blocks changed from 1048576 to 1310720

One last thing: let’s check how the free space lost the GiB we assigned:

# pvs
  PV         VG   Fmt  Attr PSize  PFree
  /dev/xvda3 vg00 lvm2 a-   87.52g 74.52g

And… back to the beginning:

# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/xvda1            3.7G  1.1G  2.7G  30% /
none                  490M  140K  490M   1% /dev
none                  500M     0  500M   0% /dev/shm
none                  500M     0  500M   0% /tmp
none                  500M   60K  500M   1% /var/run
none                  500M     0  500M   0% /var/lock
none                  500M     0  500M   0% /lib/init/rw
/dev/mapper/vg00-usr  4.0G  441M  3.6G  11% /usr
/dev/mapper/vg00-var  4.0G  216M  3.8G   6% /var
/dev/mapper/vg00-home 5.0G  4.4M  5.0G   1% /home

Success! 🙂

Sources (I didn’t really use all of them, but they might be useful to dig deeper – I tried to put the most relevant ones at the top):

Update (15 Jan 2011): checking the queries bringing people here and reproducing them in Google, I was able to find a FAQ in the UK FAQ just about that… I really wonder why the heck the 1&1 FAQ is so totally different from one country to another! Anyway, here’s the link: 1&1 Webhosting FAQ | How can I increase a logical volume?

Posted in Linux, servers.


11 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. BoMan says

    Thanks a lot for this nice tutorial. Was exactly what I needed!

  2. TJ says

    Just got a 1and1 dedicated server today. All I can say to you is Thank You, Thank You, Thank You!

  3. patheticcockroach says

    Are you saying that there is the same issue even on their dedicated servers? (not cloud/VPS, but dedicated?)

  4. TJ says

    Yep, same issue:

    # df -h
    Filesystem            Size  Used Avail Use% Mounted on
    /dev/md1              3.7G  571M  3.2G  16% /
    /dev/mapper/vg00-usr  3.9G  619M  3.1G  17% /usr
    /dev/mapper/vg00-var  3.9G  232M  3.5G   7% /var
    /dev/mapper/vg00-home 3.9G  136M  3.6G   4% /home
    none                  4.0G     0  4.0G   0% /tmp
  5. patheticcockroach says

    Wow… I would have thought you’d have the choice of a custom partitioning during installation… On the cloud I can understand because of the possibility to extend or reduce the available hard drive, but on a dedicated server that doesn’t seem to make sense…

  6. TJ says

    I agree. I’ve had dedicated servers at several places (the majority at Softlayer) and I’ve never seen such a poor system.

  7. Vincent says

    Very usefull post, it’s exactly what I was looking for, extremely tough to find good information on this Dynamic Cloud Server resizing … Was a pity to have 100GB, with no possibility to use it …

    Thks a lot

  8. patheticcockroach says

    You’re welcome. Hi from Paris, btw 🙂

  9. Jonathan says

    David, Can I pay you do to this for my 1and1 cloud server? I need to increase tmp partition to atleast 2GB

  10. patheticcockroach says

    I would if I could, but as far as I know the only folders you can resize this way are /usr, /var and /home. For some reason, /tmp as well as a few others aren’t part of the LVM volume, as you can see from df -h, and thus can’t be resized, as far as I know…
    What I could do, if you know you to configure your application to use a different temporary folder, is create another partition on the LVM volume so that you can use it as a temporary partition. But I’d prefer not to do it on a prod server, though, I’m not too confident when dealing with those things: I don’t really do this often.

  11. Anonymous says

    thanks !!! great job !!



Some HTML is OK

or, reply to this post via trackback.

Please solve the CAPTCHA below in order to fight spamWordPress CAPTCHA