[SOLVED] missing 15% of disk capacity

[mkaut]

Hello,

I have a 650 GB ext3 LVM partition with RAID 1 on. The partition is 85% full, but the system says "no space left on device" - where did the 15% go?

I ran "tune2fs -m 0 /dev/mda1", so it is not the space reserved for the root - so Nautilus reports the same free capacity as GParted now.

Some more info:
- Ubuntu 9.10 x64
- GParted says 650 GiB, 104.83 GiB free
- Nautilus says 104.8 GiB free
- The system thinks the disk is completely full - I cannot even create (touch) a new empty file

Hope someone can help me find mine 100 GB..

Michal

[JosePF]

Hi,
could you report lvdisplay and vgdisplay ouputs?

Regards

[mkaut]

I found the source of the problem, thanks to people on ubuntuforums:

$ df -i
Filesystem            Inodes   IUsed   IFree IUse% Mounted on
/dev/md1              665600  665600       0  100% /data

But I still do not know what to do about it..

Michal

[JosePF]

Hi, again

If i have understood correctly, have you a volume group that uses the RAID1 array as a physical volume and has a logical volume mounted under /data?
if it is right, you should report next commands:

mdadm --detail /dev/md1
vgdisplay
lvdisplay

If you have a LVM i think that it can be the problem of free space, not the RAID1.

Bye

[mkaut]

Hi, again

If i have understood correctly, have you a volume group that uses the RAID1 array as a physical volume and has a logical volume mounted under /data?
if it is right, you should report next commands:

mdadm --detail /dev/md1
vgdisplay
lvdisplay

If you have a LVM i think that it can be the problem of free space, not the RAID1.

Bye

I just repeat, the problem is that I have ran out of inodes on the ext3 partition - so the question is if I can increase the number of inodes without losing data?

[mkaut]

It looks one cannot change the number of inodes on an ext3 partition, so I will have to reformat. Can someone check my battle plan and fill in the blanks, please?

1. unmount the /dev/md1 drive + remove the raid (needed?, if so, how?)

2. mount /dev/sda8 (raid drive 1) and format it

3. mount /dev/sdb7 (raid drive 2) and all data from to /dev/sda8

4. format /dev/sdb7 in the same way as /dev/sda8 (needed?)

5. unmount the two drives (?)

4. create a new raid using the two drives - do I need to say which disk I want to mirrored, or will it figure out automatically?

ps1: Where does the file format "live", on /dev/sd*, or on /dev/md1?
ps2: With the danger of starting a flame war, what fs should I use for the data partition (650GB, most files about 1-10 MB (photos and music))?

Thanks

[ptkobe]

mkaut,
open a terminal and type:
$ sudo vgdisplay

Then copy the output of the command and paste it in your reply to this topic.

Do you have any question about how to do the above?

If not, do it, and then I'll see if I can answer your questions.

People are trying to help you, but you have to give them the answers they need to do so. Thus that makes sense?

Please, I'm trying to help, not to be rude or something alike.
Luis

[mkaut]

mkaut,
open a terminal and type:
$ sudo vgdisplay

Then copy the output of the command and paste it in your reply to this topic.

Do you have any question about how to do the above?

If not, do it, and then I'll see if I can answer your questions.

People are trying to help you, but you have to give them the answers they need to do so. Thus that makes sense?

Please, I'm trying to help, not to be rude or something alike.
Luis

Hello,

I still do not see how this is going to help me to find additional inodes, but anyway:
both vgdisplay and lvdisplay return nothing at all (ran with sudo). If it means anything, I did not even have lvm2 installed, had to do so just to run the commands.

One of the previous posters has also asked about the output of "mdadm --detail /dev/md1" (again ran with sudo):

/dev/md1:
        Version : 00.90
  Creation Time : Fri Nov 21 17:47:58 2008
     Raid Level : raid1
     Array Size : 681573568 (650.00 GiB 697.93 GB)
  Used Dev Size : 681573568 (650.00 GiB 697.93 GB)
   Raid Devices : 2
  Total Devices : 2
Preferred Minor : 1
    Persistence : Superblock is persistent

    Update Time : Tue Sep  7 16:14:04 2010
          State : clean
 Active Devices : 2
Working Devices : 2
 Failed Devices : 0
  Spare Devices : 0

           UUID : 0e5224d4:4903c263:fbc4a4b8:1034b584
         Events : 0.385243

    Number   Major   Minor   RaidDevice State
       0       8        8        0      active sync   /dev/sda8
       1       8       23        1      active sync   /dev/sdb7

[rcgreen]

Remember, when you reformat, there is a way to specify
a greater number of inodes than the default. This is useful
when you have lots of small files.

[ptkobe]

I still do not see how this is going to help me to find additional inodes, but anyway:
both vgdisplay and lvdisplay return nothing at all (ran with sudo). If it means anything, I did not even have lvm2 installed, had to do so just to run the commands.

You're right, I didn't get the no free inodes part. But you did start the topic saying "I have a 650 GB ext3 LVM partition with RAID 1 on", so I must be missing something else. Moving on

About the fs, I read this at inode - Wikipedia, the free encyclopedia

It is possible to use up a device's set of inodes. When this happens, new files cannot be created on the device, even though there may be free space available. For example, a mail server may have many small files that don't fill up the disk, but use many inodes to point to the numerous files.

Filesystems (such as JFS, ext4, or XFS) escape this limitation via support extents and/or dynamic inode allocation, which can 'grow' the filesystem and/or increase the number of inodes.

ps2: My opinion, I never used JFS, thik ext4 is a bit too new yet, I use XFS and like it. Also read ext4/xfs benchmarks that said ext4 is much faster, though. No flame war intended here. You may also increase the inodes for ext3 with mkfs.ext3 -i or -I options, for example.

More:
ps1: You format the /dev/mdX, sd* don't have to be formated at all, just to be partitions of type fd (linux raid autodetect)

About your plan, the best will be to have a third disk where to copy the data, right?

Without that, I don't know how to "unmount" /dev/md1 and make the information accessible through the underlying /dev/sdXx.
Your plan will stop at step 1 for me. I'll use the mdadm approach like:

use mdadm (see MANAGE MODE) to report /dev/sdbx as --fail and --remove it from array md1

--zero-superblock /dev/sdbx

CREATE a new array /dev/md2 with /dev/sdbx and a "missing" device (degraded), format it and activate. Add md2 to mdadm.conf

copy md1 content to md2

--stop md1, remove it from mdadm.conf

--zero-superblock sdax and format

--add sdax to /dev/md2.
After sync, you'll have a 650MB /dev/md2 with your data

Use the above at your own risk if you don't have a backup. Well, with a backup it will be much simpler

Regards
Luis