Eu criei uma matriz RAID10 usando 4 unidades de 75G, para criar um armazenamento de 150G.
Depois que tudo terminava (incluindo a sincronização inicial), tudo parecia bom, exceto a saída df -h
que mostrava apenas armazenamento de 73G no ponto de montagem designado.
Detalhes:
- A máquina é uma instância m1.large do Ubuntu 11.10 no Amazon EC2.
- As 4 unidades são unidades EBS, cada uma com 75G de tamanho.
- A matriz RAID10 foi criada usando o seguinte 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"
Esta é a saída que eu tenho:
*** 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
Esta é a saída relevante de df -h
:
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/mongodb_vg-mongodb_lv
73G 180M 69G 1% /ebsRaid
Esta é a saída de 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
Esta é a saída de 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:
Esta é a saída do 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:
Esta é a saída de 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
lvdisply -m
e postar por favor.Respostas:
Seu grupo de volumes não está usando a totalidade das extensões criadas para ele:
Você pode adicionar mais extensões usando o seguinte comando:
LEIA A PÁGINA DO HOMEM ANTES DE DIGITAR ESTE
Este comando estenderia o grupo de volumes para usar todas as extensões GRÁTIS restantes (você também pode selecionar menos se quiser manter algumas extensões livres). Ele utilizará as extensões no md0 para fazer isso.
Você pode redimensionar a partição online usando:
E deve dizer que está redimensionando online. Acredito que isso resolverá seu problema, mas leia as páginas de manual e entenda o que elas estão fazendo antes de tentar. Eu não sou responsável por você destruir seus discos.
Além disso, o RAID em um EBS e o lvm no topo parecem um nível desnecessário de virtualização dos discos. Você não aumentará o desempenho nem a segurança dos dados adicionando RAID extra. O LVM já faz o espelhamento / distribuição, se bem me lembro. Embora você possa executar tecnicamente o LVM no RAID no LVM no RAID ad infinitum, não tenho certeza de que você está ganhando muito com isso (embora eu esteja mais do que feliz em ser apontado como errado nisso).
fonte