Why does df shows only half the size of a RAID10 array?
I've created a RAID10 array using 4 75G drives, to create a storage of 150G.
After everything finished (including the initial syncing), everything looked good except the output of df -h which showed only 73G storage on the designated mount point.
Details:
- The machine is an m1.large Ubuntu 11.10 instance on Amazon EC2.
- The 4 drives are EBS drives, each is 75G in size.
- The RAID10 array was created using the following script:
-
#!/bin/sh
disk1="/dev/sdh1"
disk2="/dev/sdh2"
disk3="/dev/sdh3"
disk4="/dev/sdh4"
echo "*** Verifying existence of 4 volumes $disk1, $disk2, $disk3 and $disk4"
if [ -b "$disk1" -a -b "$disk2" -a -b "$disk3" -a -b "$disk4" ]; then
echo "# Found expected block devices."
else
echo "!!! Did not find expected block devices. Error."
exit -1
fi
until read -p "??? - How big (in GB) are the disks (They should be the same size)? " disk_size && [ $disk_size ]; do
echo "Please enter a disk size."
done
lv_size=$(echo "scale=2; $disk_size * 2.0" | bc)
echo "*** Assuming a per disk size of $disk_size gigs, will create a logical volume of $lv_size gigs, with $lv_size reserved for snapshots"
echo "*** Partitioning disks..."
echo "~ Partitioning $disk1"
echo ',,L' | sfdisk $disk1
echo "~ Partitioning $disk2"
echo ',,L' | sfdisk $disk2
echo "~ Partitioning $disk3"
echo ',,L' | sfdisk $disk3
echo "~ Partitioning $disk4"
echo ',,L' | sfdisk $disk4
sleep 6
echo "*** Creating /dev/md0 as a RAID 10"
/sbin/mdadm /dev/md0 --create --level=10 --raid-devices=4 $disk1 $disk2 $disk3 $disk4
echo " ~ Allocating /dev/md0 as a physical volume."
/sbin/pvcreate /dev/md0
echo " ~ Allocating a Volume Group 'mongodb_vg'"
/sbin/vgcreate -s 64M mongodb_vg /dev/md0
echo " ~ Creating a Logical Volume 'mongodb_lv'"
num_extents=$(echo "$disk_size * 1000 / 64" | bc)
/sbin/lvcreate -l $num_extents -nmongodb_lv mongodb_vg
echo " ~ Formatting the new volume (/dev/mongodb_vg/mongodb_lv) with EXT4"
/sbin/mkfs.ext4 /dev/mongodb_vg/mongodb_lv
echo " ~ Done! Go ahead and mount the new filesystem. Suggested FStab: "
echo " /dev/mongodb_vg/mongodb_lv /data ext4 defaults,noatime 0 0"
This is the output I got:
*** Verifying existence of 4 volumes /dev/xvdh1, /dev/xvdh2, /dev/xvdh3 and /dev/xvdh4
# Found expected block devices.
??? - How big (in GB) are the disks (They should be the same size)? 75
*** Assuming a per disk size of 75 gigs, will create a logical volume of 150.0 gigs, with 150.0 reserved for snapshots
*** Partitioning disks...
~ Partitioning /dev/xvdh1
Checking that no-one is using this disk right now ...
BLKRRPART: Invalid argument
OK
Disk /dev/xvdh1: 9790 cylinders, 255 heads, 63 sectors/track
sfdisk: ERROR: sector 0 does not have an msdos signature
/dev/xvdh1: unrecognized partition table type
Old situation:
No partitions found
New situation:
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0
Device Boot Start End #cyls #blocks Id System
/dev/xvdh1p1 0+ 9789 9790- 78638174+ 83 Linux
/dev/xvdh1p2 0 - 0 0 0 Empty
/dev/xvdh1p3 0 - 0 0 0 Empty
/dev/xvdh1p4 0 - 0 0 0 Empty
Warning: no primary partition is marked bootable (active)
This does not matter for LILO, but the DOS MBR will not boot this disk.
Successfully wrote the new partition table
Re-reading the partition table ...
BLKRRPART: Invalid argument
If you created or changed a DOS partition, /dev/foo7, say, then use dd(1)
to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512 count=1
(See fdisk(8).)
~ Partitioning /dev/xvdh2
Checking that no-one is using this disk right now ...
BLKRRPART: Invalid argument
OK
Disk /dev/xvdh2: 9790 cylinders, 255 heads, 63 sectors/track
sfdisk: ERROR: sector 0 does not have an msdos signature
/dev/xvdh2: unrecognized partition table type
Old situation:
No partitions found
New situation:
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0
Device Boot Start End #cyls #blocks Id System
/dev/xvdh2p1 0+ 9789 9790- 78638174+ 83 Linux
/dev/xvdh2p2 0 - 0 0 0 Empty
/dev/xvdh2p3 0 - 0 0 0 Empty
/dev/xvdh2p4 0 - 0 0 0 Empty
Warning: no primary partition is marked bootable (active)
This does not matter for LILO, but the DOS MBR will not boot this disk.
Successfully wrote the new partition table
Re-reading the partition table ...
BLKRRPART: Invalid argument
If you created or changed a DOS partition, /dev/foo7, say, then use dd(1)
to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512 count=1
(See fdisk(8).)
~ Partitioning /dev/xvdh3
Checking that no-one is using this disk right now ...
BLKRRPART: Invalid argument
OK
Disk /dev/xvdh3: 9790 cylinders, 255 heads, 63 sectors/track
sfdisk: ERROR: sector 0 does not have an msdos signature
/dev/xvdh3: unrecognized partition table type
Old situation:
No partitions found
New situation:
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0
Device Boot Start End #cyls #blocks Id System
/dev/xvdh3p1 0+ 9789 9790- 78638174+ 83 Linux
/dev/xvdh3p2 0 - 0 0 0 Empty
/dev/xvdh3p3 0 - 0 0 0 Empty
/dev/xvdh3p4 0 - 0 0 0 Empty
Warning: no primary partition is marked bootable (active)
This does not matter for LILO, but the DOS MBR will not boot this disk.
Successfully wrote the new partition table
Re-reading the partition table ...
BLKRRPART: Invalid argument
If you created or changed a DOS partition, /dev/foo7, say, then use dd(1)
to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512 count=1
(See fdisk(8).)
~ Partitioning /dev/xvdh4
Checking that no-one is using this disk right now ...
BLKRRPART: Invalid argument
OK
Disk /dev/xvdh4: 9790 cylinders, 255 heads, 63 sectors/track
sfdisk: ERROR: sector 0 does not have an msdos signature
/dev/xvdh4: unrecognized partition table type
Old situation:
No partitions found
New situation:
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0
Device Boot Start End #cyls #blocks Id System
/dev/xvdh4p1 0+ 9789 9790- 78638174+ 83 Linux
/dev/xvdh4p2 0 - 0 0 0 Empty
/dev/xvdh4p3 0 - 0 0 0 Empty
/dev/xvdh4p4 0 - 0 0 0 Empty
Warning: no primary partition is marked bootable (active)
This does not matter for LILO, but the DOS MBR will not boot this disk.
Successfully wrote the new partition table
Re-reading the partition table ...
BLKRRPART: Invalid argument
If you created or changed a DOS partition, /dev/foo7, say, then use dd(1)
to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512 count=1
(See fdisk(8).)
*** Creating /dev/md0 as a RAID 10
mdadm: partition table exists on /dev/xvdh1 but will be lost or
meaningless after creating array
mdadm: partition table exists on /dev/xvdh2 but will be lost or
meaningless after creating array
mdadm: partition table exists on /dev/xvdh3 but will be lost or
meaningless after creating array
mdadm: partition table exists on /dev/xvdh4 but will be lost or
meaningless after creating array
Continue creating array? y
mdadm: Defaulting to version 1.2 metadata
mdadm: array /dev/md0 started.
~ Allocating /dev/md0 as a physical volume.
Physical volume "/dev/md0" successfully created
~ Allocating a Volume Group 'mongodb_vg'
Volume group "mongodb_vg" successfully created
~ Creating a Logical Volume 'mongodb_lv'
Logical volume "mongodb_lv" created
~ Formatting the new volume (/dev/mongodb_vg/mongodb_lv) with EXT4
mke2fs 1.41.14 (22-Dec-2010)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
Stride=128 blocks, Stripe width=256 blocks
4800512 inodes, 19185664 blocks
959283 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=4294967296
586 block groups
32768 blocks per group, 32768 fragments per group
8192 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done
This filesystem will be automatically checked every 36 mounts or
180 days, whichever comes first. Use tune2fs -c or -i to override.
~ Done! Go ahead and mount the new filesystem. Suggested FStab:
/dev/mongodb_vg/mongodb_lv /data ext4 defaults,noatime 0 0
This is the relevant output of df -h:
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/mongodb_vg-mongodb_lv
73G 180M 69G 1% /ebsRaid
This is the output of mdadm --detail /dev/md0
/dev/md0:
Version : 1.2
Creation Time : Wed Feb 29 10:14:39 2012
Raid Level : raid10
Array Size : 157283328 (150.00 GiB 161.06 GB)
Used Dev Size : 78641664 (75.00 GiB 80.53 GB)
Raid Devices : 4
Total Devices : 4
Persistence : Superblock is persistent
Update Time : Wed Feb 29 13:21:49 2012
State : clean
Active Devices : 4
Working Devices : 4
Failed Devices : 0
Spare Devices : 0
Layout : near=2
Chunk Size : 512K
Name : my.site.com:0 (local to host my.site.com)
UUID : CENSORED
Events : 19
Number Major Minor RaidDevice State
0 202 113 0 active sync /dev/xvdh1
1 202 114 1 active sync /dev/xvdh2
2 202 115 2 active sync /dev/xvdh3
3 202 116 3 active sync /dev/xvdh4
This is the output of cat /proc/mdstat:
Personalities : [raid10]
md0 : active raid10 xvdh4[3] xvdh3[2] xvdh2[1] xvdh1[0]
157283328 blocks super 1.2 512K chunks 2 near-copies [4/4] [UUUU]
unused devices: <none>
EDIT1:
This is the output of lvdisplay -m:
--- Logical volume ---
LV Name /dev/mongodb_vg/mongodb_lv
VG Name mongodb_vg
LV UUID SEpGth-cXd3-ZFhy-XLHo-T5pV-gEd1-Tgancs
LV Write Access read/write
LV Status available
# open 1
LV Size 73.19 GiB
Current LE 1171
Segments 1
Allocation inherit
Read ahead sectors auto
- currently set to 4096
Block device 252:0
--- Segments ---
Logical extent 0 to 1170:
Type linear
Physical volume /dev/md0
Physical extents 0 to 1170
EDIT 2:
This is the output of vgdisplay:
--- Volume group ---
VG Name mongodb_vg
System ID
Format lvm2
Metadata Areas 1
Metadata Sequence No 2
VG Access read/write
VG Status resizable
MAX LV 0
Cur LV 1
Open LV 1
Max PV 0
Cur PV 1
Act PV 1
VG Size 149.94 GiB
PE Size 64.00 MiB
Total PE 2399
Alloc PE / Size 1171 / 73.19 GiB
Free PE / Size 1228 / 76.75 GiB
VG UUID CENSORED
Your Volume Group isn't using the entirety of the extents created for it:
VG Size 149.94 GiB
PE Size 64.00 MiB
Total PE 2399
Alloc PE / Size 1171 / 73.19 GiB
Free PE / Size 1228 / 76.75 GiB
You can add more extents using the following command:
lvextend -l +100%FREE /dev/mongodb_vg/mongodb_lv /dev/md0
PLEASE READ THE MAN PAGE BEFORE TYPING THIS IN
This command would extend the volume group to use all the FREE extents that are left (you can also select less if you want to keep some extents free). It will utilise the extents on md0 to do this.
You then can resize the partition online using:
resize2fs /dev/mongodb_vg/mongodb_lv
And it should say it is resizing online. I believe this will solve your issue but please read the man pages and understand what they are doing before trying. I'm not responsible for you trashing your disks.
As an aside, RAID on an EBS and then lvm on top seems like an unnecessary level of virtualising the disks. You won't enhance performance nor safety of the data by adding extra RAID in. LVM does mirroring/striping already if I remember correctly. Although you technically can run LVM on RAID on LVM on RAID ad infinitum I'm not sure you are gaining much by doing so (though I'm more than happy to be pointed out to be wrong on this).