The following is the procedure for restoring a Cpanel/WHM account from backup, with domain/account customer.biz as example:-
I. Simple Restore (From CPanel Backup)
1. Check where the account/domain is hosted, in WHMCS (Note down "Server" and "Username" in product/service item). For customer.biz:-
https://www.nocser.net/clients/admin/clientsservices.php?userid=208&id=246
( Server: bizvm.nocser.net , Username: customerbiz ).
2. Go to the hosting server's WHM --> Backup Configuration. Scroll down to "Additional Destinations", and edit the destination. Note down the "Remote Host", "Remote Account Username" and "Backup Directory". ( Example: for bizvm.nocser.net, host: 42.0.20.33, username: ftpbackup, directory: /bizvm.nocser.net )
3. SSH to the "Remote Host" via credentials in tech.nocser.net. Then
# cd ~<remote account username>
( Example: for bizvm.nocser.net, # cd ~ftpbackup )
# cd <Backup Directory without leading slash (/)>
( Example: for bizvm.nocser.net, # cd bizvm.nocser.net )
4. Find available backup snapshots for this account/domain.
# find . -iname "<Username>.*"
( Example: for customer.biz: # find . -iname "customerbiz.*" )
- Locate the latest/required available backup file
- Copy the file to the original server's home directory via scp:-
# cd <path-to-file>
# scp -vrpP 22222 <Username>.tar <hosting-server>:/home/
[ If backup file not found, the account might have been terminated/suspended for some time. Proceed to II. Restore from VM Backup ]
5. Go to hosting server's WHM --> Restore a Full Backup/cpmove File
- Enter account name ( eg customerbiz )
- Click "Restore". Wait for restoration to complete.
- If restoration failed due to existing account with same name, verify that the account is the one required for restore, then either:-
i) repeat restore using WHM --> Backup Restoration or ii) Delete the account and retry restore using above method.
II. Restore From VM Backup
- Only use this step if above procedure fails. In the case of customer.biz, the account was suspended/terminated for some time such that no cpanel backup exists.
1. Go to SolusVM and search for the hosting server's VM entry. (Example: In customer.biz case: bizvm.nocser.net)
- In VM details, take note of "ID:" ( for bizvm.nocser.net: ID: kvm000 ), and "Disk Space" ( for bizvm.nocser.net: 320GB ), and click on "Host node" ( Example: for bizvm.nocser.net: server00 )
2. In node details, click on "Auto FTP Backup". Take note the "IP or hostname", "Username", and "Directory". ( Example: for server00, IP: 10.0.20.33, Username: ftpbackup2 , Directory: /solus-node-backup ).
3. SSH to the backup server IP. If IP starts with 10.0.x.x, change the 1st part to 42 ( Example: 10.0.20.33 --> 42.0.20.33 ), using credentials found in tech.nocser.net . Then:-
# cd ~<Username>
( Example: # cd ~ftpbackup2 )
# cd <Directory without leading slash>
( Example: # cd solus-node-backup )
4. Find available backup snapshots for this VM.
# find . -iname "*<VM ID>*.bz -exec ls -lh {} \;"
( Example: for bizvm.nocser.net: # find . -iname "*kvm000*.bz -exec ls -lh {} \;" )
- Locate the latest available backup file, and take note of its size ( Example: for kvm000, size is 105GB )
# pwd
( Take note of path to backup file )
5. SSH to hosting node ( Example: for bizvm.nocser.net: server00 )
6. Check available space:-
# vgdisplay
- Take note of "VG Name" ( Example: for server00: VG Name: vg_server00 )
- Look for "Free PE / Size", make sure available size is at least the same as "Disk Space" in the VM details + backup file size.
( Example: server00: Free PE / Size 257476 / 1005.77 GiB , backup file size is 105GB, bizvm.nocser.net's Disk Space is 320GB)
7. Create temporary LV for backup file:-
# lvcreate -L<Backup File Size>G -n lv_restore_snapshot <VG Name>
( Example: for bizvm.nocser.net, in server00 run: # lvcreate -L 120G -n lv_restore_snapshot vg_server00 )
- Make sure LV is created OK:-
# lvdisplay
(Look for lv_restore_snapshot and check the details).
- Prepare ext filesystem ( if CentOS 5.x, use ext3. If CentOS 6.x, use ext4 ):-
# mkfs.ext4 /dev/<VG Name>/lv_restore_snapshot>
( Example: for server00: # mkfs.ext4 /dev/vg_server00/lv_restore_snapshot )
8. Mount temporary LV to a temporary directory:-
# mkdir /mnt/lv_restore_snapshot
# mount /dev/<VG Name>/lv_restore_snapshot /mnt/lv_restore_snapshot
( Example: for server00: # mount /dev/vg_server00/lv_restore_snapshot /mnt/lv_restore_snapshot )
- Make sure LV is mounted OK:-
# df -h
(Look for lv_restore_snapshot and check the details).
9. In node (eg server00), copy backup file from backup server to mounted temporary directory:-
# screen
# cd /mnt/lv_restore_snapshot
# scp -vrpP 22222 <Backup Server IP>:<Path to Backup File>/<Backup File> .
( Example: for bizvm.nocser.net @ server00: # scp -vrpP 22222 10.0.20.33:/export/solusvm/backups/solus-node-backup/361-kvm-kvm000-2.bz .
10. Create temporary LV for restoration:-
# lvcreate -L<VM Disk Space>G -n lv_restore <VG Name>
( Example: for bizvm.nocser.net, in server00 run: # lvcreate -L 320G -n lv_restore vg_server00 )
- Make sure LV is created OK:-
# lvdisplay
(Look for lv_restore and check the details).
11. Restore the downloaded backup file into the temporary LV (may take a long time depending on size of VM Disk Space):-
# bunzip2 -c <Backup File> | dd of=/dev/<VG Name>/lv_restore
( Example: for bizvm.nocser.net, in server00 run: # bunzip2 -c 361-kvm-kvm000-2.bz | dd of=/dev/vg_server00/lv_restore
12. Make sure partition table in restored LVM is intact:-
# fdisk -l /dev/<VG Name>/lv_restore
( Example: for server00 run: # fdisk -l /dev/vg_server00/lv_restore )
13. In order to mount the "LVM with LVM" for data extraction, use kpartx:-
# yum install kpartx
- Make partitions within LVM available
# kpartx -av /dev/<VG Name>/lv_restore
( Example: for server00 run: # kpartx -av /dev/vg_server00/lv_restore
add map vg_server00-lv_restore1 (253:10): 0 1024000 linear /dev/vg_server00/lv_restore 2048
add map vg_server00-lv_restore2 (253:11): 0 670062592 linear /dev/vg_server00/lv_restore 1026048
)
- Scan and expose all LVM volume groups within the LVM for mounting:-
# vgscan
( Example: for server00:-
Found volume group "vg_bizvm" using metadata type lvm2
Found volume group "vg_server00" using metadata type lvm2
)
- Make all LVM volumes active and ready for use
# vgchange -ay
( Example: for server00:-
3 logical volume(s) in volume group "vg_bizvm" now active
8 logical volume(s) in volume group "vg_server00" now active
)
- List available mountable LVMs inside the additional volume group found
# ls -l /dev/mapper/<Newly discovered VG Name>*
( Example: for server00, vg_bizvm is the new volume group found, to list:-
# ls -l /dev/mapper/vg_bizvm*
lrwxrwxrwx. 1 root root 8 Jan 8 10:09 /dev/mapper/vg_bizvm-lv_home -> ../dm-13
lrwxrwxrwx. 1 root root 8 Jan 8 10:09 /dev/mapper/vg_bizvm-lv_root -> ../dm-12
lrwxrwxrwx. 1 root root 8 Jan 8 10:09 /dev/mapper/vg_bizvm-lv_swap -> ../dm-14
)
14. Mount the desired partition(s) to temporary directory(ies) e.g:-
# mkdir /mnt/lv_restore
# mount /dev/mapper/<Newly discovered VG Name>-<Desired lv> /mnt/lv_restore
( Example: for server00: # mount /dev/mapper/vg_bizvm-lv_home /mnt/lv_restore )
- Make sure LV is mounted OK:-
# df -h
(Look for lv_restore and check the details).
15. Locate the desired data to be restored and manually copy to the VM.
( Example: to recover account 'customerbiz' to bizvm.nocser.net:-
- Go to WHM @ bizvm, create account customerbiz matching the details in WHMCS.
- SSH bizvm, and rename the directory: # cd /home; mv -vi customerbiz customerbiz.orig
- SSH back to server00, and copy over the files from backup:-
# cd /mnt/lv_restore/
- Copy home directory:-
# rsync -ave 'ssh -p 22222' customerbiz bizvm.nocser.net:/home/
- Copy databases to temp directory in bizvm:-
# umount /mnt/lv_restore/
# mount /dev/mapper/vg_bizvm-lv_root /mnt/lv_restore/
# cd /mnt/lv_restore
# rsync -ave 'ssh -p 22222' var/lib/mysql/customerbiz* bizvm.nocser.net:/root/
- SSH bizvm, and correct permissions:-
# cd /home
# chown -R customerbiz:customerbiz customerbiz
# chgrp -R mail customerbiz/etc
# chgrp nobody customerbiz/public_html
- Test browse http://customer.biz and correct database permissions following config
files in /home/customerbiz/public_html:-
- In WHM -> List Accounts -> customerbiz --> CPanel --> MySQL Database Wizard
- Recreate each restored database which was copied in /root/, following
credentials in config files in /home/customerbiz/public_html (look for config files
or # grep -r "customerbiz_" /home/customerbiz/public_html and open the file ).
16. Clean up when done (Important! Otherwise will occupy space in server)
# umount /mnt/lv_restore
vgchange -an <Newly discovered VG Name>
( Example: for server00: # vgchange -an vg_bizvm )
# kpartx -d /dev/<VG Name>/lv_restore
( Example: for server00: # kpartx -d /dev/vg_server00/lv_restore )
# lvremove <VG Name>/lv_restore
( Example: for server00: # lvremove vg_server00/lv_restore )
# lvremove <VG Name>/lv_restore_snapshot
( Example: for server00: # lvremove vg_server00/lv_restore_snapshot )
- Verify temporary LVs no longer exist.
# lvdisplay
- Verify volume group free space back to normal
# vgdisplay
DONE.
APPENDIX
1. Sometimes you may encounter the following error during bunzip2 of backup file:-
[root@server00 lv_restore_snapshot]# bunzip2 -c 361-kvm-kvm000-1.bz | dd of=/dev/vg_server00/lv_restore
bunzip2: Compressed file ends unexpectedly;
perhaps it is corrupted? *Possible* reason follows.
bunzip2: Success
Input file = 361-kvm-kvm000-1.bz, output file = (stdout)
It is possible that the compressed file(s) have become corrupted.
You can use the -tvv option to test integrity of such files.
You can use the `bzip2recover' program to attempt to recover
data from undamaged sections of corrupted files.
350676796+1 records in
350676796+1 records out
179546520000 bytes (180 GB) copied, 18613.1 s, 9.6 MB/s
[root@server00 lv_restore_snapshot]#
- Ignore the error and resume with the steps following it. You might be able
to still mount the needed partition. (If cannot mount later, use another
backup file)
