meta data for this page
Duply
Installation
sudo apt-get install duply python-paramiko trickle
Configuration
Create backup profile:
duply gitlab create
Profile file $HOME/.duply/gitlab/conf
was created.
IMPORTANT Copy the whole profile folder after the first backup to a safe place. It contains everything needed to restore your backups. You will need it if you have to restore the backup from another system (e.g. after a system crash). Keep access to these files restricted as they contain all information (gpg data, ftp data) to access and modify your backups.
Repeat this step after all configuration changes. Some configuration options are crucial for restoration.
Generate random password:
openssl rand -base64 20
- | ~/.duply/gitlab/conf
#GPG_KEY= GPG_PW='<generated passsword>'
Configure backup section:
- | ~/.duply/gitlab/conf
# Paramiko SSH is very CPU consuming #TARGET='scp://gitlabbackup@192.168.0.230//mnt/backup/gitlabbackup' TARGET='sftp://gitlabbackup@192.168.0.230//mnt/backup/gitlabbackup' # Limit network speed DUPL_PRECMD="trickle -s -u 1500 -d 256" SOURCE='/' MAX_AGE=6M MAX_FULL_BACKUPS=2 MAX_FULLS_WITH_INCRS=2 MAX_FULLBKP_AGE=3M DUPL_PARAMS="$DUPL_PARAMS --full-if-older-than $MAX_FULLBKP_AGE " VOLSIZE=256 DUPL_PARAMS="$DUPL_PARAMS --volsize $VOLSIZE " VERBOSITY=4 TEMP_DIR=/tmp # Specify different id_rsa file: DUPL_PARAMS="$DUPL_PARAMS --ssh-options=-oIdentityFile='/root/.duply/gitlab/id_rsa' " DUPL_PARAMS="$DUPL_PARAMS --no-compression " # Do not use GZip to compress files on remote system. DUPL_PARAMS="$DUPL_PARAMS --ssh-options=-carcfour128 " # light cipher for NAS DUPL_PARAMS="$DUPL_PARAMS --ssh-options=-oCompression=no " # disable ssh compression DUPL_PARAMS="$DUPL_PARAMS --asynchronous-upload " # uploads in background
- ~/.duply/gitlab/exclude
+ /etc/gitlab + /opt + /home + /root - **
Note: duply dosn't make any cleanup or deletions during backup
action. So destination storage can be full very quickly.
To perform maintenance of old backup accordign to MAX_AGE
, MAX_FULL_BACKUPS
, MAX_FULLS_WITH_INCRS
and MAX_FULLBKP_AGE
parameters it is need to call
duply with purge
and cleanup
commands. See cron script example below.
Example options:
- MAX_FULL_BACKUPS=2
- MAX_FULLS_WITH_INCRS=1
Will keep 2 full backup sets, but only one with increments (last one).
Sometimes it is good to check whether incremental backups are meaningful (it depends on type of data stored). If command
duply gitlab status
shows that number of volume with each increment is similar to full backup, then making increments has no sense, ane time to keep increments can be short e.g. MAX_FULLBKP_AGE=7D
Usage
Start the backup
sudo duply gitlab backup --progress
duply gitlab status
duply gitlab list
duply gitlab verify # long operation
cron script
#!/bin/bash -ue set -o pipefail trap "banner error; echo LINE: $LINENO" ERR duply gitlab backup duply gitlab purge --force # list outdated backup archives and delete them duply gitlab-to-grinnux purgeIncr --force duply gitlab-to-grinnux purgeFull --force duply gitlab cleanup --extra-clean --force > /dev/null # list broken backup files and delete them banner ALL OK
shell function
#!/bin/bash -ueE set -o pipefail trap "banner error; echo LINE: $LINENO" ERR run_duply() { echo "=====================================================" duply ${1} backup echo "=====================================================" duply ${1} cleanup --extra-clean --force duply ${1} purge --force duply ${1} purgeIncr --force duply ${1} purgeFull --force echo "=====================================================" duply ${1} cleanup --extra-clean --force > /dev/null echo "=====================================================" banner ${1} OK }
SFTP and rbash
- /etc/passwd
mybackup:x:1002:1002:Backup,,,:/home/mybackup:/bin/rbash
- /etc/ssh/sshd_config
Subsystem sftp internal-sftp
Verbosity
VERBOSITY=5
Lots of info
Ignoring incremental Backupset (start_time: Fri Aug 5 13:49:11 2016; needed: Fri Jun 17 14:20:07 2016) Ignoring incremental Backupset (start_time: Fri Aug 5 13:49:11 2016; needed: Thu Jul 21 14:33:52 2016) Added incremental Backupset (start_time: Fri Aug 5 13:49:11 2016 / end_time: Wed Aug 10 14:54:49 2016) Ignoring incremental Backupset (start_time: Wed Aug 10 14:54:49 2016; needed: Fri Jun 17 14:20:07 2016)
VERBOSITY=4
Usefull report:
-------------[ Backup Statistics ]-------------- StartTime 1479517583.39 (Sat Nov 19 01:06:23 2016) EndTime 1479517751.96 (Sat Nov 19 01:09:11 2016) ElapsedTime 168.57 (2 minutes 48.57 seconds) SourceFiles 41 SourceFileSize 13621422991 (12.7 GB) NewFiles 1 NewFileSize 4096 (4.00 KB) DeletedFiles 0 ChangedFiles 1 ChangedFileSize 13621360640 (12.7 GB) ChangedDeltaSize 0 (0 bytes) DeltaEntries 2 RawDeltaSize 6101758 (5.82 MB) TotalDestinationSizeChange 5214178 (4.97 MB) Errors 0 -------------------------------------------------
Issues
no acceptable kex algorithm
ssh: Exception: Incompatible ssh peer (no acceptable kex algorithm)
Python paramiko module needs upgrade
apt-get install python-pip python-dev python-cffi libffi-dev build-essential pip install --ugprade cffi pip install pycparser==2.13 pip install --ugprade cryptography
To solve error “AssertionError: sorry, but this version only supports 100 named groups” please install
pip install pycparser==2.13
pip install --ugprade paramiko
can't be deleted
duply mybackup purge --force
Last full backup date: Wed May 24 01:11:54 2017 There are backup set(s) at time(s): Thu Nov 24 01:05:26 2016 Fri Nov 25 01:09:43 2016 Sat Nov 26 01:10:50 2016 Which can't be deleted because newer sets depend on them. No old backup sets found, nothing deleted.
Solution is to run:
duply mybackup purgeIncr --force
duply mybackup purgeFull --force