Benutzer-Werkzeuge
know-how:backup
Inhaltsverzeichnis
Windows
- Windows Sicherung muss installiert sein
- Generiert VHDX Datein die gemountet werden können
Externer SAMBA Server rotierend
- backup-generic.bat
- Bare Metal Backup falls die Maschine selbst etwas hat
- Achtung unter BACKUP_TARGET müssen sich dann wie in diesem Fall die Verzeichnisse 0,1 befinden (wegen Module Operation der Anzahl der Wochen die gespeichert werden sollen)
setlocal
set LOGFILE=F:\Backup-Logging.txt
set G=%temp%\getWeek.vbs
set WEEKS=2
set USERNAME=USERNAME_SAMBA
set PASSWORD=PASSWORD_SAMBA
set BACKUP_TARGET=\\backup\backup_serverxy\bare_metal
>%G% echo WScript.Echo Datepart("ww",WScript.Arguments(0),2)
for /f %%i in ('cscript //nologo %G% %date%') do set calWeek=%%i
set /a DIRECTORY=%calweek%%%WEEKS%
date /t >> %LOGFILE%
wbadmin start backup -quiet -vssCopy -user:"%USERNAME%" -password:"%PASSWORD%" -allCritical -backupTarget:%BACKUP_TARGET%%DIRECTORY% >> %LOGFILE% 2>&1
"Externes" Laufwerk rotierend
- backup-storagexy.bat
- Achtung Hier wird das Laufwerk G:\ auf F:\ gesichert, wieder müssen die Verzeichnisse 0,1 unter BACKUP_TARGET existieren
setlocal
set LOGFILE=F:\Backup-Logging-StorageXY.txt
set G=%temp%\getWeek.vbs
set WEEKS=2
set BACKUP_TARGET=\\localhost\F$\Backups\STORAGEXY\
>%G% echo WScript.Echo Datepart("ww",WScript.Arguments(0),2)
for /f %%i in ('cscript //nologo %G% %date%') do set calWeek=%%i
set /a DIRECTORY=%calweek%%%WEEKS%
echo "----BEGIN BACKUP----" >> %LOGFILE% 2>&1
date /t >> %LOGFILE%
wbadmin start backup -vssCopy -quiet -include:G:\ -backupTarget:%BACKUP_TARGET%%DIRECTORY% >> %LOGFILE% 2>&1
echo "----END BACKUP----" >> %LOGFILE%
Externe Festplatten
- Verschiedene Bat Dateien mit Laufwerksangaben für verschiedene externe Festplatten
- gelöscht werden muss manuell
- zB: G:\Backups\RECHNERNAME\DATUM
- backup-extern-g.bat
@echo off & setlocal
set TARGET=G$
set NAME=%COMPUTERNAME%
for /f %%i in ('date /T') do Set DIRECTORY=%%i
if not exist "\\localhost\%TARGET%\Backups" goto fehler
mkdir \\localhost\%TARGET%\Backups\%NAME%\%DIRECTORY%\
wbadmin start backup -quiet -vssFull -allCritical -backupTarget:\\localhost\%TARGET%\Backups\%NAME%\%DIRECTORY%\
pause
exit
:fehler
echo Konnte Festplatte %TARGET% nicht finden ggf. \\localhost\%TARGET%\Backups Verzeichnis anlegen
pause
exit
Hyper-V Guests auf Festplatte rotierend
- backup-hyper-v.bat
setlocal
set LOGFILE=PFAD\Backup-Logging-Guests.txt
set G=%temp%\getWeek.vbs
set WEEKS=4
set BACKUP_TARGET=\\localhost\F$\Guests\
>%G% echo WScript.Echo Datepart("ww",WScript.Arguments(0),2)
for /f %%i in ('cscript //nologo %G% %date%') do set calWeek=%%i
set /a DIRECTORY=%calweek%%%WEEKS%
echo "----BEGIN BACKUP----" >> %LOGFILE% 2>&1
date /t >> %LOGFILE%
wbadmin start backup -quiet -hyperv:"GUESTNAME1,GUESTNAME2,GUESTNAME3" -backupTarget:%BACKUP_TARGET%%DIRECTORY% >> %LOGFILE% 2>&1
echo "----END BACKUP----" >> %LOGFILE%
Duplicati 2 - verschlüsseltes Cloud Backup
- Anfoderungen
- Deduplizierendes, verschlüsseltes Backup , das in der „Cloud“ abgelegt werden kann in diesem Fall ins One-Drive Business von einem User Account der als „Backup“ User angelegt wurde
- Lösungsmöglichkeit
- Duplicati 2: https://www.duplicati.com/download
- Achtung BETA
- One-Drive Einstellungen
- Achtung v2 auswählen und mit Graph API Key erstellen - bei >5000 Dateien in einem Ordner kommt es sonst zu „Limits“ die schwer zu debuggen sind
- Getestet auf Windows 2016 Standard und Duplicati 2.0.4.23 - 64bit , Backup enthält > 1 Million Dateien
- Wiederherstellung von einzelnen Dateien hat beim Testen funktioniert war jedoch sehr langsam
Urbackup
- Fehler Referenzmaschine war „Hyper-V“ Client und wurde auf physikalische Maschine migriert - Fehlermeldung am Client „Hyper-V IC Software Shadow Copy Provider ist noch installiert
2019-11-06 06:31:52: ERROR: Not found
2019-11-06 06:31:56: ERROR: backupcom->AddToSnapshotSet(&(Server->ConvertToWchar(selected_vols[i])[0]), GUID_NULL, &additional_refs[i].volid) failed. VSS error code VSS_E_UNEXPECTED_PROVIDER_ERROR
2019-11-06 06:31:56: ERROR: VSS provider information:
2019-11-06 06:31:56: ERROR: Anbietername: "Hyper-V IC Software Shadow Copy Provider"
2019-11-06 06:31:56: ERROR: Anbietertyp: Software
2019-11-06 06:31:56: ERROR: Anbieterkennung: {74600e39-7dc5-4567-a03b-f091d6c7b092}
2019-11-06 06:31:56: ERROR: Version: 1.0.0.0
2019-11-06 06:31:56: ERROR: Anbietername: "Microsoft Software Shadow Copy provider 1.0"
2019-11-06 06:31:56: ERROR: Anbietertyp: System
2019-11-06 06:31:56: ERROR: Anbieterkennung: {b5946137-7b9f-4925-af80-51abd60b20d5}
2019-11-06 06:31:56: ERROR: Version: 1.0.0.7
2019-11-06 06:31:56: ERROR: Creating shadowcopy of "C:" failed.
2019-11-06 06:31:56: ERROR: Creating shadow copy failed. See client log file for details.
2019-11-06 07:12:01: ERROR: Not found
2019-11-06 07:12:03: ERROR: backupcom->AddToSnapshotSet(&(Server->ConvertToWchar(selected_vols[i])[0]), GUID_NULL, &additional_refs[i].volid) failed. VSS error code VSS_E_UNEXPECTED_PROVIDER_ERROR
2019-11-06 07:12:03: ERROR: VSS provider information:
2019-11-06 07:12:03: ERROR: Anbietername: "Hyper-V IC Software Shadow Copy Provider"
2019-11-06 07:12:03: ERROR: Anbietertyp: Software
2019-11-06 07:12:03: ERROR: Anbieterkennung: {74600e39-7dc5-4567-a03b-f091d6c7b092}
2019-11-06 07:12:03: ERROR: Version: 1.0.0.0
2019-11-06 07:12:03: ERROR: Anbietername: "Microsoft Software Shadow Copy provider 1.0"
2019-11-06 07:12:03: ERROR: Anbietertyp: System
2019-11-06 07:12:03: ERROR: Anbieterkennung: {b5946137-7b9f-4925-af80-51abd60b20d5}
2019-11-06 07:12:03: ERROR: Version: 1.0.0.7
2019-11-06 07:12:03: ERROR: Creating shadowcopy of "C:" failed.
2019-11-06 07:12:03: ERROR: Creating shadow copy failed. See client log file for details.
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\VSS\Providers\{74600e39-7dc5-4567-a03b-f091d6c7b092}]
"vssadmin list providers" sollte nach einem Reboot den Hyper V Provider nicht mehr anzeigen
- Bei der ISO Datei für den Restore wird Kernel 4.19 (Debian) verwendet
- Eintrag für PXE BIOS Boot über TFTP (https://urbackup.atlassian.net/wiki/spaces/US/pages/1441795/Restoring+Images+from+PXE) - dann könnte auch anderer Kernel verwendet werden hinsichtlich Treiberunterstützung
LABEL UrBackup Restore menu label Urbackup Restore Backup kernel urbackup/live/vmlinuz append initrd=urbackup/live/initrd.img boot=live config username=urbackup toram noswap fetch=tftp://IP_ADRESSE_TFTP_SERVER/urbackup/live/filesystem.squashfs
HP Aruba
- Für Cronjob siehe „HP Pro Curve“
- SSH Server Keys müssen bereits „accepted“ worden sein / sonst hängt er bei der Accept Key Frage
- backup-aruba.expect
#!/usr/bin/expect -f set timeout 60 log_user 0 spawn ssh USERNAME@[lindex $argv 0] expect "password:" send "PASSWORD\r" #2021-02-08 cc: think wisely if you want to send password as argument to call #send "[lindex $argv 1]\r" expect "continue" send "\r" expect "#" send "terminal length 1000\r" expect "#" log_user 1 send "show config\r" expect "#" send "exit\r" expect ">" send "exit\r" expect "?" send "y\r" expect eof
HPE Comware
- Für Cronjob siehe „HP Pro Curve“
- SSH Server Keys müssen bereits „accepted“ worden sein / sonst hängt er bei der Accept Key Frage
- backup-hp-comware-switches.expect
#!/usr/bin/expect -f set timeout 60 spawn ssh admin@[lindex $argv 0] expect "password:" send "PASSWORD\r" #2021-02-08 cc: think wisely if you want to send password as argument to call #send "[lindex $argv 1]\r" expect ">" send "xtd-cli-mode\r" expect "]:" send "Y\r" expect "Password:" send "foes-bent-pile-atom-ship\r" expect ">" send "screen-length disable\r" expect ">" send "show current-configuration\r" expect ">" send "quit\r"
HP Pro Curve
- Achtung Sicherheitsimplikationen bei telnet beachten (Übertragung des passworts über plaintext)
- /etc/cron.d/backup-switches
# Every day @02:00 a.m. SHELL=/bin/bash 00 2 * * * root switches="IP_SWITCH_1 IP_SWITCH_2 IP_SWITCH_3 IP_SWITCH_4"; for i in $switches ; do /usr/local/bin/get-switch-config $i > /var/backups/switches-configs/$i.conf ; done
- Dependency: expect
- /usr/local/bin/get-switch-config
#!/usr/bin/expect -f
# To avoid empty lines, 'nonewline' flag is used
set ip_address [lindex $argv 0];
set prompt "#"
set username "USERNAME\r"
set password "PASSWORD\r"
log_user 0
spawn telnet $ip_address
expect "Username:"
send $username
expect "Password:"
send $password
expect $prompt
send "terminal length 1000\r"
expect $prompt
set timeout 120
log_user 1
send "show config\r"
expect $prompt
set timeout 20
send "exit\r"
expect ">"
send "exit\r"
expect "?"
send "y\r"
expect eof
Linux
rsnapshot
- Erstellung von Prüfsummen zu den Dateien die gebackuped wurden
- rsnapshot.conf - Beispiele für die Verwendung von integ.sh
... cmd_postexec /usr/local/sbin/integ.sh /mnt/external-backup/ebackup1/weekly.0 c ; /bin/mount -o remount,ro /mnt/external-backup ...
- rsnapshot.conf - Beispiele für die Verwendung von backup-lvm-drive.sh
... backup_script /usr/local/sbin/backup-lvm-drive.sh /dev/vgroup/lvmname imagename.img.lzo ./pfad/ ...
- backup-lvm-drive.sh
- Kann als backup_script ausgeführt werden zB: wenn auf einem Hostsystem virtualisiert über kvm eine „docker“ blackbox läuft und lvm
#!/bin/bash
export PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
function bailout ()
{
echo -e "$1"
exit 2
}
function usage ()
{
echo -e "$1"
exit 1
}
# LVM PATH / generate LVM snapshot name suffix snapshot / compress devices / output PATH
ARGUMENTS=$#
SIZE_SNAPSHOT="20G"
BUFFER="10M"
which lvcreate > /dev/null || bailout "lvcreate not found"
which lvremove > /dev/null || bailout "lvremove not found"
which lzop > /dev/null || bailout "lzop not found - apt-get install lzop"
[ $ARGUMENTS != 2 ] && bailout "Usage: $0 PATH_TO_LVM_VOLUME BACKUP_FILE"
LVM_PATH="$1"
BACKUP_FILE="$2"
[ ! -r $LVM_PATH ] && bailout "FAIL: Cannot read PATH_TO_LVM VOLUME"
[ -r $BACKUP_FILE ] && bailout "FAIL: $BACKUP_FILE is already here aborting"
LVM_NAME=$(echo ${LVM_PATH##*/})
LVM_NAME_SNAPSHOT="$LVM_NAME""_snapshot"
LVM_PATH_SNAPSHOT="${LVM_PATH%%$LVM_NAME}$LVM_NAME_SNAPSHOT"
[ -r $LVM_PATH_SNAPSHOT ] && bailout "FAIL: Snapshot: $LVM_PATH_SNAPSHOT already here aborting"
lvcreate -L $SIZE_SNAPSHOT -s -n $LVM_NAME_SNAPSHOT $LVM_PATH > /dev/null || bailout "FAIL: Could not create Snapshot: $LVM_NAME_SNAPSHOT"
if [ $BACKUP_FILE == '-' ]
then
dd if=$LVM_PATH_SNAPSHOT bs=$BUFFER 2>/dev/null | lzop -
else
dd if=$LVM_PATH_SNAPSHOT bs=$BUFFER 2>/dev/null | lzop > $BACKUP_FILE
fi
lvremove -q -f $LVM_PATH_SNAPSHOT > /dev/null
exit 0
- backup-lvm-files.sh
- Um einen LVM Snapshot read only zu mounten und dann per RSYNC mit einem Verzeichnis zu syncen / kann mit rsnapshot backup_exec kombiniert werden
#!/bin/bash
function bailout ()
{
echo -e "$1"
[ -z "$2" ] || lvremove -q -f "$2"
exit 2
}
function usage ()
{
echo -e "$1"
exit 1
}
RSYNC_DEFAULT_OPTIONS="--dry-run -a --delete --numeric-ids "
RSYNC_DEFAULT_OPTIONS_ARRAY=($RSYNC_DEFAULT_OPTIONS)
ARGUMENTS=$#
SIZE_SNAPSHOT="10G"
BUFFER="10M"
TMP_MOUNT_POINT="/mnt/tmp"
which lvcreate > /dev/null || bailout "lvcreate not found"
which lvremove > /dev/null || bailout "lvremove not found"
which rsync > /dev/null || bailout "rsync not found"
which kpartx > /dev/null || bailout "kpartx not found"
which mount > /dev/null || bailout "mount not found"
which mountpoint > /dev/null || bailout "mountpoint not found"
if [ $ARGUMENTS != 2 ] && [ $ARGUMENTS != 3 ]
then
bailout "Usage: $0 PATH_TO_LVM_VOLUME BACKUP_DIRECTORY [ \""RSYNC_OPTIONS\"" ]"
fi
LVM_PATH="$1"
BACKUP_DIRECTORY="$2"
RSYNC_CUSTOM_OPTIONS="$3"
[ -n "$3" ] && RSYNC_DEFAULT_OPTIONS_ARRAY+=($RSYNC_CUSTOM_OPTIONS)
[ ! -r $LVM_PATH ] && bailout "FAIL: Cannot read $LVM_PATH VOLUME"
[ ! -d $BACKUP_DIRECTORY ] && bailout "FAIL: $BACKUP_DIRECTORY is already here aborting"
[ ! -d $TMP_MOUNT_POINT ] && bailout "FAIL: $TMP_MOUNT_POINT directory not found"
mountpoint -q $TMP_MOUNT_POINT && bailout "FAIL: $TMP_MOUNT_POINT is already mounted FAIL"
LVM_NAME=$(echo ${LVM_PATH##*/})
LVM_NAME_SNAPSHOT="$LVM_NAME""_snapshot"
LVM_PATH_SNAPSHOT="${LVM_PATH%%$LVM_NAME}$LVM_NAME_SNAPSHOT"
[ -r $LVM_PATH_SNAPSHOT ] && bailout "FAIL: Snapshot: $LVM_PATH_SNAPSHOT already here aborting"
lvcreate -L $SIZE_SNAPSHOT -s -n $LVM_NAME_SNAPSHOT $LVM_PATH > /dev/null || bailout "FAIL: Could not create Snapshot: $LVM_NAME_SNAPSHOT"
[ $? == "0" ] || bailout "FAIL: $LVM_PATH_SNAPSHOT could not be created"
MOUNT_PATH="$LVM_PATH_SNAPSHOT"
DEVICE_MAPPER_ENTRY=$(kpartx -av $LVM_PATH_SNAPSHOT)
RETURN_SUCCESS=$?
[ $? != "0" ] && bailout "FAIL: Kpartx did not return 0 - error" "$LVM_PATH_SNAPSHOT"
DEVICE_MAPPER_ENTRY=$(echo $DEVICE_MAPPER_ENTRY | cut -f 3 -d" ")
[ -z $DEVICE_MAPPER_ENTRY ] || MOUNT_PATH="/dev/mapper/""$DEVICE_MAPPER_ENTRY"
mount -o ro "$MOUNT_PATH" "$TMP_MOUNT_POINT" || bailout "FAIL: could not mount $MOUNT_PATH to $TMP_MOUNT_POINT" "$LVM_PATH_SNAPSHOT"
rsync "${RSYNC_DEFAULT_OPTIONS_ARRAY[@]}" "$TMP_MOUNT_POINT/" "$BACKUP_DIRECTORY/"
RETURN_RSYNC=$?
umount $TMP_MOUNT_POINT
kpartx -dv $LVM_PATH_SNAPSHOT > /dev/null
lvremove -q -f $LVM_PATH_SNAPSHOT > /dev/null
exit $RETURN_RSYNC
- integ.sh
- Um nach dem Backup Prüfsummen aller Dateien zu erstellen
#!/bin/bash
MACHINE="MEIN-SERVER-NAME"
function usage {
echo "Usage: $0 directory_root [c|v]"
echo "c...create hashes"
echo "v...veriy hashes"
exit 0
}
echo $MACHINE
function verifyDirectoryHashes {
echo "verify"
cd $DIRECTORY_ROOT
[[ ! -f hashes.sha1 ]] && echo "Hashes: $DIRECTORY_ROOT/hashes.sha1 not found" && exit 2
date1=$(date -u +"%s")
sha1sum --quiet -c hashes.sha1
retCode=$?
date2=$(date -u +"%s")
diff=$(($date2-$date1))
amount=$(wc -l hashes.sha1 | cut -d " " -f 1)
echo "$(($diff / 60)) minutes and $(($diff % 60)) seconds elapsed."
echo "Hashes verified: $amount"
echo "done"
exit $retCode
}
function createDirectoryHashes {
echo "create hashes"
cd $DIRECTORY_ROOT
echo -n > hashes.sha1
date1=$(date -u +"%s")
find ./ ! -name "*.sha1" -type f -exec sha1sum {} >> hashes.sha1 \;
date2=$(date -u +"%s")
diff=$(($date2-$date1))
amount=$(wc -l hashes.sha1 | cut -d " " -f 1)
echo "$(($diff / 60)) minutes and $(($diff % 60)) seconds elapsed."
echo "Hashes created: $amount"
echo "done"
exit 0
}
#Check parameters
DIRECTORY_ROOT="$1"
ACTION="$2"
[[ ! -d $DIRECTORY_ROOT ]] && echo "Cannot read directory: $DIRECTORY_ROOT" && usage
[[ $ACTION != "c" && $ACTION != "v" ]] && echo "Either verify or create" && usage
[[ $ACTION == "c" ]] && createDirectoryHashes
[[ $ACTION == "v" ]] && verifyDirectoryHashes
- rsnapshot mit btrfs
rsnapshot.conf ... cmd_cp /usr/local/sbin/rsnapshot_btrfs_cp.sh # uncomment this to use the rm program instead of the built-in perl routine. # cmd_rm /usr/local/sbin/rsnapshot_btrfs_rm.sh ...
- rsnapshot_btrfs_cp.sh
#!/bin/bash # Arg 1: -al # Arg 2: /path/daily.0 # Arg 3: /path/daily.1 btrfs subvolume snapshot -r $2 $3 >/dev/null
- rsnapshot_btrfs_rm.sh
#!/bin/bash
# Arg 1: -rf
# Arg 2: /path/daily.5/
# echo 1: $1 2: $@
# Try to delete the given path with btrfs subvolume delete first
# if this fails fall back to normal rm
if [ "$1" = "-rf" -a "$3" = "" ]; then
# "trying to delete with btrfs"
btrfs subvolume show "$2" &>/dev/null && btrfs subvolume delete "$2" &>/dev/null && exit 0
[[ -d "$2" ]] && rm -rf "$2" && exit 0
exit 2
fi
exit 2
- rsnapshot mit borg cron job
- In der Konfiguration der rsnapshot Backups existiert nur mehr daily.0 d.h. retain daily 1 und beide Backup Jobs (customers und pannoniait) müssen bereits abgeschlossen sein
30 06 * * 1-5 root ( [[ ! -f /var/run/rsnapshot-customers.pid ]] && [[ ! -f /var/run/rsnapshot-pannoniait.pid ]] && borg create -x --numeric-owner --compression lz4 /mnt/storage/backups/borg/::d.$(date +"\%F") /mnt/storage/backups/customers/daily.0/ /mnt/storage/backups/pannoniait/daily.0/ && borg prune -P d -d 5 /mnt/storage/backups/borg ) || echo "Could not perform daily borg backup" 30 06 * * 6 root ( [[ ! -f /var/run/rsnapshot-customers.pid ]] && [[ ! -f /var/run/rsnapshot-pannoniait.pid ]] && borg create -x --numeric-owner --compression lz4 /mnt/storage/backups/borg/::w.$(date +"\%F") /mnt/storage/backups/customers/daily.0/ /mnt/storage/backups/pannoniait/daily.0/ && borg prune -P w -w 4 /mnt/storage/backups/borg ) || echo "Could not perform weekly borg backup" 30 07 1 * * root ( [[ ! -f /var/run/rsnapshot-customers.pid ]] && [[ ! -f /var/run/rsnapshot-pannoniait.pid ]] && borg create -x --numeric-owner --compression lz4 /mnt/storage/backups/borg/::m.$(date +"\%F") /mnt/storage/backups/customers/daily.0/ /mnt/storage/backups/pannoniait/daily.0/ && borg prune -P m -m 2 /mnt/stroage/backups/borg ) || echo "Could not perform monthly borg backup
borg - generell
- Backup Tests mit borg - Achtung unverschlüsselt da Testgerät ohnehin vollverschlüsselt ist
- Durchgeführt auf Kali Linux - Debian Testing
Borg Befehle: 1) Repository initialisieren d.h. Verzeichnis initialisieren wo das Backup landen soll e.g. Encryption modes Hash/MAC Not encrypted no auth Not encrypted, but authenticated Encrypted (AEAD w/ AES) and authenticated SHA-256 none authenticated repokey keyfile BLAKE2b n/a authenticated-blake2 repokey-blake2 keyfile-blake2 ( encryption -> 'none', 'keyfile', 'repokey', 'authenticated', 'keyfile-blake2', 'repokey-blake2', 'authenticated-blake2' ) root@mrChief:/home# borg init --encryption none /mnt/backup/mrChief/ root@mrChief:/home# echo $? 0 2) Backup erstellen Exclude mir alle ISO Dateien unter /home/urnilxfgbez/ Nicht über das Dateisystem rausschießen deshalb / und /boot einzeln Nummerische Ids der Ownership speichern borg create --stats --progress --one-file-system --numeric-owner --exclude /swap.img --exclude '/home/*/*.iso' --exclude '/home/*/*.ISO' --compression lz4 /mnt/backup/mrChief/::$(date +"%T.%F") / /boot 3) Verfügbare Backups anzeigen (borg list) root@mrChief:/home# borg list /mnt/backup/mrChief/ 12:13:01.2019-10-16 Wed, 2019-10-16 12:13:02 [6e550028349bdde6f22ab513a83169ebf6ef87026af80e5598f7e9c82dff4229] 4) Verfügbares Backup mounten und zugänglich machen root@mrChief:/mnt# borg mount /mnt/backup/mrChief::12:13:01.2019-10-16 /mnt/tmp root@mrChief:/mnt# borg umount /mnt/tmp root@mrChief:/mnt# ls -al /mnt/tmp total 8 drwxr-xr-x 2 root root 4096 Jan 23 2019 . drwxr-xr-x 14 root root 4096 Oct 16 09:47 .. 5) Backup Policy definieren f. Aufbewahrungsfristen (borg prune) Letzte 4 Versionen behalten root@mrChief:/mnt# borg prune --keep-last 4 /mnt/backup/mrChief/
borg - hetzner storage
1. key verteilen
https://wiki.hetzner.de/index.php/Backup_Space_SSH_Keys
root@mrStorage:~# echo -e "mkdir .ssh \n chmod 700 .ssh \n put /root/.ssh/id_rsa.pub .ssh/authorized_keys \n chmod 600 .ssh/authorized_keys" | sftp u12345678@u12345678.your-storagebox.de
The authenticity of host 'u12345678.your-storagebox.de (176.9.161.59)' can't be established.
RSA key fingerprint is SHA256:EMlfI8GsRIfpVkoW1H2u0zYVpFGKkIMKHFZIRkf2ioI.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added 'u12345678.your-storagebox.de,176.9.161.59' (RSA) to the list of known hosts.
u12345678@u12345678.your-storagebox.de's password:
Connected to u12345678@u12345678.your-storagebox.de.
sftp> mkdir .ssh
sftp> chmod 700 .ssh
Changing mode on /.ssh
sftp> put /root/.ssh/id_rsa.pub .ssh/authorized_keys
Uploading /root/.ssh/id_rsa.pub to /.ssh/authorized_keys
/root/.ssh/id_rsa.pub 100% 740 27.0KB/s 00:00
sftp> chmod 600 .ssh/authorized_keys
Changing mode on /.ssh/authorized_keys
2. Repo initialisieren
borg init --encryption=keyfile ssh://u12345678@u12345678.your-storagebox.de:23/./borg/
Keymaterial befindet sich ausschließlich auf "mrStorage" dem Server der das Backup in die Hetzner CLoud durchführt
3. Backup Keys exportieren - es wird Keyfile und Passphrase benötigt - backup des keyfiles
borg key export ssh://u12345678@u12345678.your-storagebox.de:23/./borg/ /tmp/borg_keyfile
3. Common Commands
root@mrStorage:~# borg create --compression lz4 --list --dry-run ssh://u12345678@u12345678.your-storagebox.de:23/./borg/::{now} /root/fw.log.2.gz
- /root/fw.log.2.gz
root@mrStorage:~# borg create --compression lz4 --list ssh://u12345678@u12345678.your-storagebox.de:23/./borg/::{now} /root/fw.log.2.gz
Enter passphrase for key /root/.config/borg/keys/u12345678_your_storagebox_de__borg:
A /root/fw.log.2.gz
root@mrStorage:~# borg create --compression lz4 --list --dry-run ssh://u12345678@u12345678.your-storagebox.de:23/./borg/::{now} /root/fw.log.2.gz
- /root/fw.log.2.gz
root@mrStorage:~# borg info ssh://u12345678@u12345678.your-storagebox.de:23/./borg/
Enter passphrase for key /root/.config/borg/keys/u12345678_your_storagebox_de__borg:
Location: ssh://u12345678@u12345678.your-storagebox.de:23/./borg
Encrypted: Yes (key file)
Key file: /root/.config/borg/keys/u12345678_your_storagebox_de__borg
------------------------------------------------------------------------------
Original size Compressed size Deduplicated size
All archives: 4.48 MB 4.50 MB 4.50 MB
Unique chunks Total chunks
Chunk index: 5 5
root@mrStorage:~# borg list ssh://u12345678@u12345678.your-storagebox.de:23/./borg/
Enter passphrase for key /root/.config/borg/keys/u12345678_your_storagebox_de__borg:
2020-03-05T23:35:02 Thu, 2020-03-05 23:35:11 [0fe96fe6a40b0eb5db5955affad70a570e3a37cb3d3dd9f6e9a3ba0a538eafc2]
root@mrStorage:~# borg mount ssh://u12345678@u12345678.your-storagebox.de:23/./borg/ /mnt/tmp
Enter passphrase for key /root/.config/borg/keys/u12345678_your_storagebox_de__borg:
- /usr/local/bin/hetzner_borg_backup.sh
#!/bin/bash
export BORG_PASSPHRASE="PASSPHRASE"
REPOSITORY="ssh://u12345678@u12345678.your-storagebox.de:23/./borg/"
DIRECTORIES="/mnt/storage/"
HOST=$(hostname --fqdn)
START_TIME=$(date +%Y.%m.%d-%H.%M.%S)
date1=$(date -u +"%s")
echo -e "Program: $0 \nstarted at: $START_TIME\nBacking Up: $DIRECTORIES\nHost: $HOST"
#2020-04-08 cc: Achtung sonst konsumiert er den gesamten Upload und die Leitung steht daher: Rate limit ~6 Mbit upload / bei 10Mbit verfügbarem Upload
borg create --compression lz4 --stats --remote-ratelimit 700 --exclude '*/.snapshots/*' $REPOSITORY::{now} $DIRECTORIES
[[ $? == "0" ]] && borg prune --stats --keep-last 6 $REPOSITORY
END_TIME=$(date +%Y.%m.%d-%H.%M.%S)
date2=$(date -u +"%s")
diff=$(($date2-$date1))
echo "$(($diff / 60)) minutes and $(($diff % 60)) seconds elapsed."
echo "$(($diff / 86400 )) days elapsed. "
echo -e "Program ended successfully : $0 \nended at: $END_TIME\n"
triggered
- Anforderungen:
- Es sollen Backups durchgeführt werden „Event“ basierend d.h. wenn eine bestimmte Festplatte angesteckt wird
- Getestet auf Debian Buster
- Die Seagate Platte wird sporadisch angesteckt (siehe lsusb - ID) :
root@mrGodfather:/mnt/ebackup/ebackup1# lsusb .... .... Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub Bus 005 Device 008: ID 0bc2:61b7 Seagate RSS LLC Maxtor M3 Portable .....
- Udev Trigger für diese spezielle Festplatte
root@mrGodfather:/mnt/ebackup/ebackup1# cat /etc/udev/rules.d/30-ebackup.rules
ACTION=="add",KERNEL=="sd?[0-9]",ATTRS{idVendor}=="0bc2", ATTRS{idProduct}=="61b7", RUN+="/usr/local/sbin/backup_wrapper_systemd_ebackup1.sh"
- Workaround damit „Backup“ Skript systemd entzogen wird (dort gibts ein definiertes timeout für Skript) und at übergeben
root@mrGodfather:/mnt/ebackup/ebackup1# cat /usr/local/sbin/backup_wrapper_systemd_ebackup1.sh #!/bin/bash echo "/usr/local/sbin/backup_ebackup1.sh" | at now exit 0
- Backup Skript , das crypsetup Device erstellt und Backups durchführt (Achtung btrfs Dateisystem)
root@mrGodfather:/mnt/ebackup/ebackup1# cat /usr/local/sbin/backup_ebackup1.sh
#!/bin/bash
#1st step check if device is here
#2nd step check if already mounted
#3rd step mount using cryptsetup key
#4th step remember time backup of all hosts using rsync
#5th step take snapshot
#6th sync data umount drive
export PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
KEY_LOCATION="LOCATION_TO_KEYFILE"
MAPPER_NAME="ebackup1"
SUBVOLUME_NAME="ebackup1"
MOUNT_POINT="/mnt/ebackup"
HARDDISK_ID="/dev/disk/by-id/usb-Seagate_M3_Portable_NA844AER-0:0-part1"
MAIL_NOTIFICATION="EMAIL_ADDRESS"
RSYNC_STANDARD="rsync -a --timeout=300 --acls --delete --numeric-ids --exclude mnt/backup/ --exclude proc/ --exclude sys/ --exclude dev/ --exclude backup/ --exclude mnt/storage/backups/pannoniait/ --exclude mnt/storage/backups/customers/ --exclude .snapshots/ "
MACHINES="server1 server2 server3 server4"
BACKUP_ID=$(date +%s)
START_TIME=$(date +%Y.%m.%d-%H.%M.%S)
function bailout ()
{
echo -e "$1"
exit 2
}
function notify ()
{
echo -e "$1"
}
notify "Program: $0 \nstarted at: $START_TIME\nMachines to backup: $MACHINES mrGodfather\nDestination Hardware: $HARDDISK_ID\nBackup ID: $BACKUP_ID\nDestination MountPoint: $MOUNT_POINT\nKeyfile: $KEY_LOCATION"
which cryptsetup > /dev/null || bailout "FAIL: cryptsetup not found!"
which mail > /dev/null || ( echo "FAIL: mail not found!" && exit 2 )
which rsync > /dev/null || bailout "FAIL: rsync not found!"
which btrfs > /dev/null || bailout "FAIL: btrfs not found!"
[ ! -r $HARDDISK_ID ] && bailout "FAIL: Physical Harddisk: $HARDDISK_ID not found\nBackup ID: $BACKUP_ID"
[ ! -r $KEY_LOCATION ] && bailout "FAIL: Key not found: $KEY_LOCATION\nBackup ID: $BACKUP_ID"
[ ! -d $MOUNT_POINT ] && bailout "FAIL: Mount point not found: $MOUNT_POINT\nBackup ID: $BACKUP_ID"
[ -r "/dev/mapper/$MAPPER_NAME" ] && bailout "FAIL: Crypto Mapper: $MAPPER_NAME found close manualy\nBackup ID: $BACKUP_ID"
mountpoint -q $MOUNT_POINT && bailout "FAIL: Already mounted: $MOUNT_POINT\nBackup ID: $BACKUP_ID"
cryptsetup luksOpen $HARDDISK_ID $MAPPER_NAME --key-file $KEY_LOCATION || bailout "FAIL: Could not open cryptsetup device $HARDDISK_ID with $KEY_LOCATION\nBackup ID: $BACKUP_ID"
mount -o rw,acl,noexec /dev/mapper/$MAPPER_NAME $MOUNT_POINT || bailout "FAIL: could not mount: /dev/mapper/$MAPPER_NAME on $MOUNT_POINT\nBackup ID: $BACKUP_ID"
[ ! -d "$MOUNT_POINT/$SUBVOLUME_NAME" ] && bailout "FAIL: Directory not found on Mountpoint: $MOUNT_POINT/$SUBVOLUME_NAME\nBackup ID: $BACKUP_ID"
for machine in $MACHINES
do
$RSYNC_STANDARD root@$machine:/ $MOUNT_POINT/$SUBVOLUME_NAME/$machine/
done
btrfs subvolume snapshot -r $MOUNT_POINT/$SUBVOLUME_NAME $MOUNT_POINT/$SUBVOLUME_NAME/.snapshots/@GMT_$START_TIME
END_TIME=$(date +%Y.%m.%d-%H.%M.%S)
btrfs filesystem sync $MOUNT_POINT || bailout "FAIL: could not sync btrfs filesystem\nBackup ID: $BACKUP_ID"
/usr/local/sbin/clearLastSnapshot.sh $MOUNT_POINT ebackup1 5 || bailout "FAIL: could not clearLastSnapshot\nCheck Manualyy!\mBackup ID: $BACKUP_ID"
sync
sleep 60
umount $MOUNT_POINT || bailout "FAIL: Could not unmount filesystem\nBackup ID: $BACKUP_ID"
cryptsetup luksClose $MAPPER_NAME || bailout "FAIL: Could not close cryptsetup Handle $MAPPER_NAME\nBackup ID: $BACKUP_ID"
notify "Program ended successfully : $0 \nended at: $END_TIME\nBackup ID: $BACKUP_ID"
exit 0
- Skript um nur eine bestimmte Anzahl an Snapshots auf dem Device zu halten
root@mrGodfather:/mnt/ebackup/ebackup1# cat /usr/local/sbin/clearLastSnapshot.sh
#!/bin/bash
function usage
{
echo "Usage Keep this Nr of Snapshots: $0 LocalMountPoint LocalSubvolumeName DesiredSnapshotCount"
echo "Usage Show Nr of Snapshots: $0 LocalMountPoint LocalSubvolumeName"
echo "Usage: e.g. $0 /mnt/storage daten 3"
exit 1
}
LOCAL_MOUNT_POINT=$1
LOCAL_SUBVOLUME=$2
DESIRED_SNAPSHOTS=$3
[[ $# != 3 && $# != 2 ]] && usage
[[ ! -d $LOCAL_MOUNT_POINT ]] && echo "Couldn't validate local btrfs subvolume mountpoint: $LOCAL_MOUNT_POINT" && exit 2
CURRENT_NR_SNAPSHOTS=$(btrfs subvolume list $LOCAL_MOUNT_POINT/$LOCAL_SUBVOLUME/.snapshots -r -o --sort=+gen | wc -l )
[[ "$CURRENT_NR_SNAPSHOTS" == 0 ]] && echo "Couldn't aquire number of snapshots from $LOCAL_MOUNT_POINT/$LOCAL_SUBVOLUME/.snapshots" && exit 2
[[ $# == 2 ]] && echo -e "Mount Point: $LOCAL_MOUNT_POINT\nSubvolume: $LOCAL_SUBVOLUME\nCurrent Snapshots: $CURRENT_NR_SNAPSHOTS" && exit 0
REGEX_NUMBER='^[0-9]+$'
[[ ! $DESIRED_SNAPSHOTS =~ $REGEX_NUMBER ]] && echo "That's not a valid number: $NR_SNAPSHOTS" && exit 2
[[ $(($CURRENT_NR_SNAPSHOTS-$DESIRED_SNAPSHOTS)) -le 0 ]] && echo -e "Deletion not needed\nMount Point: $LOCAL_MOUNT_POINT\nSubvolume: $LOCAL_SUBVOLUME\nCurrent Snapshots: $CURRENT_NR_SNAPSHOTS\nDesired: $DESIRED_SNAPSHOTS" && exit 0
NR_SNAPSHOTS_REMOVE=$(($CURRENT_NR_SNAPSHOTS-$DESIRED_SNAPSHOTS))
CURRENT_SNAPSHOTS=$(btrfs subvolume list $LOCAL_MOUNT_POINT/$LOCAL_SUBVOLUME/.snapshots -r -o --sort=+gen | head -n $NR_SNAPSHOTS_REMOVE | cut -d' ' -f 9 )
for snap in $CURRENT_SNAPSHOTS
do
btrfs subvolume delete $LOCAL_MOUNT_POINT/$snap
done
btrfs filesystem sync $LOCAL_MOUNT_POINT
etc
- rsync für blockdevices Kopieren “–copy-devices„
- Update
rsync (3.2.0-1) unstable; urgency=low This latest release changed two parameters which used to be present on the Debian packaging of rsync as upstream now integrated the patches. Previous parameter: --copy-devices: write to devices as files (implies --inplace) Is now called: --write-devices
root@mrChief:/home/urnilxfgbez# rsync -v --progress --block-size=131072 --no-whole-file --checksum --copy-devices /dev/mmcblk0 /tmp/foo
- Interessantes zu der Option und beim Kopieren von/auf lokalem Dateisystem: https://superuser.com/questions/234273/why-doest-rsync-use-delta-transfer-for-local-files
- Er scheint zuerst die Hashes vom Ziel zu berechnen , danach von der Quelle (debugged mit iostat auf dem jeweiligen Device)
root@mrChief:/home/urnilxfgbez# rsync -vv --no-whole-file --copy-devices /dev/mmcblk0 /tmp/foo delta-transmission enabled mmcblk0 total: matches=126225 hash_hits=126225 false_alarms=0 data=0 sent 505,008 bytes received 1,009,926 bytes 5,439.62 bytes/sec total size is 0 speedup is 0.00 root@mrChief:/home/urnilxfgbez# rsync -vv --no-whole-file --copy-devices root@localhost:/dev/mmcblk0 /tmp/foo opening connection using: ssh -l root localhost rsync --server --sender -vve.LsfxC --copy-devices . /dev/mmcblk0 (11 args) delta-transmission enabled mmcblk0 total: matches=126225 hash_hits=126225 false_alarms=0 data=0 sent 1,009,903 bytes received 505,075 bytes 5,439.78 bytes/sec
- rsync - Ausgabe ob etwas geändert werden würde - dry run und count
- rsync -iaun –delete foo1/ foo2/ | wc -l
0
know-how/backup.txt · Zuletzt geändert: 2021/02/10 11:51 von cc
Seiten-Werkzeuge
