Lennart's weblog

Open source, computers, Africa and other more (or less) interesting stuff.

Tag: script (page 1 of 2)

Moving annual backups from an external disk with Ext4 to an external disk with ZFS

For a few years I have used the Christmas holidays to create a full
backup of my /home on an external hard disk. For that I used a
Bash script around rsync that uses hard links to keep the used disk
space under control. Each backup was saved in a directory named with
the date of the backup. POSIX ACLs were also backed up.

Since last year’s backup I have moved to ZFS (using ZFS on Linux
with Ubuntu 14.04
) as filesystem for /home (and others). Since ZFS
makes checksums of data and metadata it has the possibility to
detect corrupted files (and if the data is redundant it can also fix
them). This is a feature I’d like to have for my backups as
well: I’d rather know it when corruption occurs than live in
ignorance.

So the plan is to move the old backups from the external disk to the
ZFS pool in my server. and instead of using hard links I’ll transfer
the backups in order from old to new to the ZFS pool making a
snapshot for each. Additionally I will also turn on compression
(using the lz4 algorithm). Once that is done I will reformat the
external drive and create a ZFS pool called “JaarlijkseBackupPool” on
it (jaarlijks means annual in Dutch).

The old situation

In the current/old situation, this is how much disk space is used
on the external disk (with and without taking the hard links into
account):

$ sudo du -csh /mnt/JaarlijkseBackups/*
102G    /mnt/JaarlijkseBackups/2010-11-28
121G    /mnt/JaarlijkseBackups/2013-02-04
101G    /mnt/JaarlijkseBackups/2013-12-23
324G    total
$ sudo du -clsh /mnt/JaarlijkseBackups/*
102G    /mnt/JaarlijkseBackups/2010-11-28
193G    /mnt/JaarlijkseBackups/2013-02-04
255G    /mnt/JaarlijkseBackups/2013-12-23
549G    total

Copying the data from the Ext4 disk to a temporary ZFS filesystem on my server

The ZFS pool in my server is called storage. In order to save the
POSIX ACLs of the Ext4 system, they need to be enabled when
creating the ZFS filesystem as well. Setting xattr=sa means the
ACLS are stored more efficiently (although this option is not
compatible with other ZFS implementations at this time, so if I
would try to import the ZFS pool in FreeBSD for example, that
information would be inaccessible).

$ zfs create storage/JaarlijkseBackupsOrganized \
      -o compression=lz4 \
      -o acltype=posixacl \
      -o xattr=sa
$ sudo rsync -ahPAXHS --numeric-ids \
     /storage/JaarlijkseBackups/2010-11-28/ \
     /storage/JaarlijkseBackupsOrganized
$ zfs snapshot storage/JaarlijkseBackupsOrganized@2010-11-28

Followed by the same for the same rsync and zfs snapshot
commands for the other two dates.
Once that is finished, this is the status of that ZFS FS:

$ zfs list -r -t all storage/JaarlijkseBackupsOrganized
NAME                                            USED  AVAIL  REFER  MOUNTPOINT
storage/JaarlijkseBackupsOrganized              275G   438G   272G  /storage/JaarlijkseBackupsOrganized
storage/JaarlijkseBackupsOrganized@2010-11-28  1,03G      -  88,9G  -
storage/JaarlijkseBackupsOrganized@2013-02-04  2,33G      -   196G  -
storage/JaarlijkseBackupsOrganized@2013-12-23      0      -   272G  -
$ zfs get -r -t all compressratio storage/JaarlijkseBackupsOrganized
NAME                                           PROPERTY       VALUE  SOURCE
storage/JaarlijkseBackupsOrganized             compressratio  1.13x  -
storage/JaarlijkseBackupsOrganized@2010-11-28  compressratio  1.19x  -
storage/JaarlijkseBackupsOrganized@2013-02-04  compressratio  1.14x  -
storage/JaarlijkseBackupsOrganized@2013-12-23  compressratio  1.12x  -

Partitioning the external disk

The external disk is as 1TB Samsung SATA 3Gbps SpinPoint F2 EcoGreen disk
(type HD103SI, serial number: S1VSJD6ZB02657). The disk uses 512B
sectors:

sudo hdparm -I /dev/sdf |grep Sector
     Logical/Physical Sector size:           512 bytes

Before using it with ZFS, it needs to be partitioned. I used
parted:

$ parted /dev/sdf
GNU Parted 2.3
Using /dev/sdf
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) p
Model: ATA SAMSUNG HD103SI (scsi)
Disk /dev/sdf: 1000GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos

Number  Start   End     Size    Type     File system  Flags
 1      1049kB  1000GB  1000GB  primary  ext4

(parted) mklabel
New disk label type? gpt
(parted) u
Unit?  [compact]? MB
(parted) p
Model: ATA SAMSUNG HD103SI (scsi)
Disk /dev/sdf: 1000205MB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start  End  Size  File system  Name  Flags

(parted) mkpart
Partition name?  []? JaarlijkseBackups-HD103SI-S1VSJD6ZB02657
File system type?  [ext2]? zfs
Start? 1M
End? 1000204M
(parted) p
Model: ATA SAMSUNG HD103SI (scsi)
Disk /dev/sdf: 1000205MB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End        Size       File system  Name                                  Flags
 1      1,05MB  1000204MB  1000203MB  ext4         JaarlijkseBackups-HD103SI-S1VSJD6ZB0

(parted) q

This removes the old partition table and creates a new GPT
partition table (which allows naming partitions). Next I set the
units to MB so I can leave 1MB at the beginning and end of the
partition (can be helpful when importing this pool in
e.g. FreeBSD). The disk also shows up in /dev/disk/by=partlabel
now.

Creating the new ZFS pool

$ zpool create -o ashift=9 JaarlijkseBackupPool \
    /dev/disk/by-partlabel/JaarlijkseBackups-HD103SI-S1VSJD6ZB0
$ zpool status JaarlijkseBackupPool
  pool: JaarlijkseBackupPool
 state: ONLINE
  scan: none requested
config:

        NAME                                    STATE     READ WRITE CKSUM
        JaarlijkseBackupPool                    ONLINE       0     0     0
          JaarlijkseBackups-HD103SI-S1VSJD6ZB0  ONLINE       0     0     0

errors: No known data errors

Migrating the data

Now that the new ZFS pool and filesystem are all in place, it is
time to move the backups to the new place, starting with the oldest
backup. The -R option also make sure the attributes like
compression and xattr are transferred to the new FS. The
following commands send each snapshot to the new pool (the -n
option of zfs receive is for doing a dry run, just to show how it
works). After the first snapshot is sent, the other two are sent
using the -i option to zfs send so that only the incremental
differences between the snapshots are sent.

$ zfs send -vR storage/JaarlijkseBackupsOrganized@2010-11-28 | \
      zfs receive -Fvu JaarlijkseBackupPool/oldRsyncBackups
$ zfs send -vR -i storage/JaarlijkseBackupsOrganized@2010-11-28 \
    storage/JaarlijkseBackupsOrganized@2013-02-04 | \
    zfs receive -Fvu JaarlijkseBackupPool/oldRsyncBackups
$ zfs send -vR -i storage/JaarlijkseBackupsOrganized@2013-02-04 \
      storage/JaarlijkseBackupsOrganized@2013-12-23 | \
      zfs receive -Fvu JaarlijkseBackupPool/oldRsyncBackups -n
send from @2013-02-04 to storage/JaarlijkseBackupsOrganized@2013-12-23 estimated size is 84,3G
total estimated size is 84,3G
TIME        SENT   SNAPSHOT
would receive incremental stream of storage/JaarlijkseBackupsOrganized@2013-12-23 into JaarlijkseBackupPool@2013-12-23
14:09:16   4,22M   storage/JaarlijkseBackupsOrganized@2013-12-23
14:09:17   8,46M   storage/JaarlijkseBackupsOrganized@2013-12-23
14:09:18   18,4M   storage/JaarlijkseBackupsOrganized@2013-12-23
14:09:19   24,8M   storage/JaarlijkseBackupsOrganized@2013-12-23
^C
$ zfs send -vR -i  storage/JaarlijkseBackupsOrganized@2013-02-04 \
      storage/JaarlijkseBackupsOrganized@2013-12-23 | \
      zfs receive -Fvu JaarlijkseBackupPool/oldRsyncBackups

Add this year’s backup

At first I tried to add the new backups also to the
oldRsyncBackups FS, but that didn’t work (at least not with an
incremental backup), so I ended up making a new backup. The extra
cost in disk space is not a real problem. Disk space is rather
cheap and the current configuration will last me at least one more
year. So after creating a snapshot called 2014-12-26 of my
/home I ran:

   $ zfs send -v  storage/home@2014-12-26 | \
      zfs receive -Fu JaarlijkseBackupPool/home
$ zfs list -r -t all JaarlijkseBackupPool
NAME                                              USED  AVAIL  REFER  MOUNTPOINT
JaarlijkseBackupPool                              581G   332G    30K  /JaarlijkseBackupPool
JaarlijkseBackupPool/home                         311G   332G   311G  /JaarlijkseBackupPool/home
JaarlijkseBackupPool/home@2014-12-26             51,2M      -   311G  -
JaarlijkseBackupPool/oldRsyncBackups              271G   332G   267G  /JaarlijkseBackupPool/oldRsyncBackups
JaarlijkseBackupPool/oldRsyncBackups@2010-11-28   974M      -  87,1G  -
JaarlijkseBackupPool/oldRsyncBackups@2013-02-04  2,23G      -   193G  -
JaarlijkseBackupPool/oldRsyncBackups@2013-12-23      0      -   267G  -
$ zfs get -r compressratio JaarlijkseBackupPool
NAME                                             PROPERTY       VALUE  SOURCE
JaarlijkseBackupPool                             compressratio  1.15x  -
JaarlijkseBackupPool/home                        compressratio  1.17x  -
JaarlijkseBackupPool/home@2014-12-26             compressratio  1.17x  -
JaarlijkseBackupPool/oldRsyncBackups             compressratio  1.13x  -
JaarlijkseBackupPool/oldRsyncBackups@2010-11-28  compressratio  1.19x  -
JaarlijkseBackupPool/oldRsyncBackups@2013-02-04  compressratio  1.14x  -
JaarlijkseBackupPool/oldRsyncBackups@2013-12-23  compressratio  1.12x  -

Finishing up

In order to be able to disconnect the external drive without
damaging the filesystems use

zpool export JaarlijkseBackupPool

Later, the drive/pool can be imported using the zpool import
command.

Now that the migration is done, the intermediate filesystem
(including the snapshots) can also be removed:

zfs destroy -r storage/JaarlijkseBackupsOrganized

For reference: the old rsync script

#!/bin/sh
#
# Time-stamp: <2013-02-04 16:48:31 (root)>
# This scripts helps me create my annual backups to an external hard
# disk. The script uses rsync's hard link option to make hard links to
# the previous backups for files that haven't changed. It makes the
# backup based on an LVM snapshot it creates of the LV that contains
# the /home partition.
# This script needs to be run as root.
 
today=`date +%F`
olddate="2013-02-04"
 
srcdir="/mnt/backupsrc/"
destdir="/mnt/backupdest/JaarlijkseBackups/$today"
prevdir="/mnt/backupdest/JaarlijkseBackups/$olddate"
 
# LVM options
VG=raid5vg
LV=home
 
# rstnc options
options="-ahPAXHS --numeric-ids"
exclusions="--exclude 'lost+found/'"
#  --exclude '*/.thumbnails'"
# exclusions="$exclusions --exclude '*/.gvfs/'"
# exclusions="$exclusions --exclude '*/.cache/' --exclude '**/Cache'"
# exclusions="$exclusions --exclude '*/.recycle/'"
 
# Check to see if the previous backup directory exists
if [ ! -d $prevdir ]; then
    echo "Error: The directory with the previous back up ($prevdir) doesn't exist" 1>&2
    exit 1
fi
 
# Make a snapshot of the home LV that we can backup
lvcreate -L15G -s -n snap$LV /dev/$VG/$LV
mount /dev/$VG/snap$LV $srcdir
 
 
# Start the backup, first a dry-run, then the full one
rsynccommand="rsync $options $exclusions --link-dest=$prevdir $srcdir $destdir"
 
$rsynccommand -n
 
# Wait for user input
echo "This was a dry run. Press a key to continue with the real stuff or"
echo "hit Ctrl-c to abort."
read dummy
 
$rsynccommand

Using rsync to backup a ZFS file system to a remote Synology Diskstation

Some time ago I moved from using LVM to using ZFS on my home server. This meant I also had to change the backup script I used to make backups on a remote Synology Diskstation. Below is the updated script. I also updated it such that it now needs a single command line argument: the hostname of the Diskstation to backup to (because I now have two Diskstations at different locations). If you want to run this script from cron you should set up key-based SSH login (see also here and here).

#!/bin/bash
#
# This script makes a backup of my home dirs to a Synology DiskStation at
# another location. I use ZFS for my /home, so I make a snapshot first and
# backup from there.
#
# This script requires that the first command line argument is the
# host name of the remote backup server (the Synology NAS). It also
# assumes that the location of the backups is the same on each
# remote backup server.
#
# Time-stamp: <2014-10-27 11:35:39 (L.C. Karssen)>
# This script it licensed under the GNU GPLv3.
 
set -u
 
if [ ${#} -lt 1 ]; then
    echo -n "ERROR: Please specify a host name as first command" 1>&2
    echo " line option" 1>&2
    exit -1
fi
 
###############################
# Some settings
###############################
# Options for the remote (Synology) backup destination
DESTHOST=$1
DESTUSER=root
DESTPATH=/volume1/Backups/
DEST=${DESTUSER}@${DESTHOST}:${DESTPATH}
 
# Options for the client (the data to be backed up)
# ZFS options
ZFS_POOL=storage
ZFS_DATASET=home
ZFS_SNAPSHOT=rsync_snapshot
SNAPDIR="/home/.zfs/snapshot/$ZFS_SNAPSHOT"
 
# Backup source path. Don't forget to have trailing / otherwise
# rsync's --delete option won't work
SRC=${SNAPDIR}/
 
# rsync options
OPTIONS="--delete -azvhHSP --numeric-ids --stats"
OPTIONS="$OPTIONS --timeout=60 --delete-excluded"
OPTIONS="$OPTIONS --skip-compress=gz/jpg/mp[34]/7z/bz2/ace/avi/deb/gpg/iso/jpeg/lz/lzma/lzo/mov/ogg/png/rar/CR2/JPG/MOV"
EXCLUSIONS="--exclude lost+found --exclude .thumbnails --exclude .gvfs"
EXCLUSIONS="$EXCLUSIONS --exclude .cache --exclude Cache"
EXCLUSIONS="$EXCLUSIONS --exclude .local/share/Trash"
EXCLUSIONS="$EXCLUSIONS --exclude home/lennart/tmp/Downloads/*.iso"
EXCLUSIONS="$EXCLUSIONS --exclude home/lennart/.recycle"
EXCLUSIONS="$EXCLUSIONS --exclude _dev_dvb_adapter0_Philips_TDA10023_DVB*"
 
 
 
###############################
# The real work
###############################
 
# Create the ZFS snapshot
if [ -d $SNAPDIR ]; then
    # If the directory exists, another backup process may be running
    echo "Directory $SNAPDIR already exists! Is another backup still running?"
    exit -1
else
    # Let's make snapshots
    zfs snapshot $ZFS_POOL/$ZFS_DATASET@$ZFS_SNAPSHOT
fi
 
 
# Do the actual backup
rsync -e 'ssh' $OPTIONS $EXCLUSIONS $SRC $DEST
 
# Remove the ZFS snapshot
if [ -d $SNAPDIR ]; then
    zfs destroy $ZFS_POOL/$ZFS_DATASET@$ZFS_SNAPSHOT
else
    echo "$SNAPDIR does not exist!" 1>&2
    exit 2
fi
 
exit 0

Using ‘expect’ to distribute files among users

I’m currently teaching at the Summmer School in Statistical Omics in Split, Croatia. A great experience!

Because of the computations involved in the project work, we have access to a server. However, since the machine is part of a university cluster, I haven’t been given full root permissions (in fact, I’m only allowed to use sudo to install packages).

Now, the problem I had to solve was that I needed to distribute a certain file (.Renviron) to each student’s home directory. Normally I’d use sudo to do that, but the admin hadn’t allowed me to use cp via sudo. Furtunately, I had a list of user names and passwords for the students (because I had to distribute those), so I thought I’d use su - to change to each student’s account and copy the file, something along the lines of

echo PASSWORD | su -

and then loop over each account. Unfortunately, while testing the script I found out it wouldn’t work since su complained:

su: must be run from a terminal

Then I remembered the expect tool, which executes commands based on what it ‘sees’ on the command line. In this case I wanted it to enter the password at su‘s prompt. This is the expect script I came up with, it accepts two command line arguments, the user name and the password:

#!/usr/bin/expect -f
 
set user [lindex $argv 0]
set pass [lindex $argv 1]
 
spawn su - $user
expect "Password: "
send "$pass\r"
expect "$ "
send "cp -i /common/WORK/school/lennart/.Renviron .\r"
expect "$ "
send "ls -l .Renviron\r"
expect "$ "
send "exit\r"

The script was wrapped in the Bash script that I had already written:

#!/bin/bash
#
# This script is used to copy files from this directory to the
# home directories of the users listed in $USERFILE.
 
USERFILE=accounts.txt
SRCFILE=/common/WORK/school/lennart/.Renviron
 
while read user passw; do
    ./copy_file_to_users.expect $user $passw
done < $USERFILE

Fixing colours in git output after upgrading to Ubuntu 14.04

After upgrading my Ubuntu 13.10 installation to 14.04, I noticed that the output of several git commands (e.g. git diff and git log) didn’t show colours as they used to, but showed ESC[ ANSI codes instead.
A quick internet search lead to this post on unix.stackexchange.com where the LESS environment variable was ‘blamed’. Indeed, I have my LESS variable (re-) defined in my .bashrc and .zshrc files.

The solution was to add -R to the environment variable, which allows raw control characters to be displayed. I now have the environment variable defined as:

LESS='--quiet -X -F -R'

Permantly ban an IP address with fail2ban

Over the last few days I noticed in my logwatch e-mails that one IP address kept trying to log in to my server, even though it was blocked regularly by fail2ban.

Here’s a post that explains how to simply add a list of IP addresses to block permanently. There’s only one catch: the listing provided there contains an error, the word <name> is missing in the iptables command, probably due to HTML conversion. This is the correct line to be insterted into the actionstart section of /etc/fail2ban/action.d/iptables-multiport.conf:

cat /etc/fail2ban/ip.blacklist | while read IP; do iptables -I fail2ban-<name> 1 -s $IP -j DROP; done

Use the following command to check if the IP address is indeed banned:

$ sudo iptables  -L fail2ban-ssh
Chain fail2ban-ssh (1 references)
target     prot opt source               destination         
DROP       all  --  192.168.20.25        anywhere            
RETURN     all  --  anywhere             anywhere 

Replacing a character in a Bash variable name

Today I needed to replace a : in a bunch of file names with a -, so I wanted to write a Bash for-loop to do just that. I vaguely remembered that you can do character replacements within variables, but couldn’t remember the details.

This is how it’s done:

for filename in *; do
    mv "$filename" "${filename/:/-}"
done

I put the variables in double quotes, because the file names contained spaces.

Doing a quick fixed-effects meta-analysis using the Rmeta package

This is a quick example of how to do a fixed meta-analysis using the R package Rmeta, just so I dont have to look it up again next time:

## Create data frame containing betas and standard errors
df <- data.frame()
df <- rbind(df, c(2., 0.2))
df <- rbind(df, c(2.5, 0.4))
df <- rbind(df, c(2.2, 0.2))
 
## Add study names
df <- cbind(df, c("study 1", "study 2", "study 3"))
 
colnames(df) <- c("beta", "se_beta", "name") 
 
## Do the meta-analysis 
ms <- meta.summaries(df$beta, df$se_beta, names=df$name)
 
## Add some colors
mc <- meta.colors(summary="darkgreen", zero="red")
 
## Make a forest plot
plot(ms, xlab=expression(beta ~ " (mmol/l)"), 
     ylab="Study", colors=mc, zero=2.6)

The resulting plot looks like this:
Forest plot of fake data

Exit a Bash script if an error occurs

Last week I found out that a Bash script I wrote to do some data QC gave me a false sense of security: a script continues even if one (or more) of the statements in the scripts fails (with an exit status not equal to 0). It turned out that for some of the data sets the QC wasn’t done correctly because I didn’t check the exit status after each step.

My first thought was: oh boy, that means I have to check $? for every step. That means a lot of repetitive code to write! Luckily my colleague came with the answer: add

set -e

at the top of you Bash script and the script will fail if one of its statements fails (for the fine print see the top answer in this is StackOverflow post).

Speeding up grep when looking in large files

In my line of work it is not uncommon to have to find out whether a given term is present in a long list. Say, for example you need to look up whether a set of, say 10, SNPs is present in a (possibly annotated) list of SNPs present on a genotyping array (having for example 240k SNPs).
My first instinct in such cases is to use grep, and it’s a good instinct that has served me well over the years.

Recently we had a case that involved quite some larger files. We needed to see whether a set of genomic positions was present in a genome-wide list of such positions. Of course we split the files up per chromosome, but still this took ~ 24 hours for a chromosome when using

grep -w -f short_list long_file > results

I was convinced this could be done faster and googled a bit, read the grep man page to find out that the -F option of grep ensures that the search string is not seen as a (regexp) pattern, but as fixed. This meant an enormous speed improvement. Instead of having to wait for 24 hours we got the output in under a minute!

I did a quick performance comparison: looking up ten items in a ~415MB file with 247,871 rows and 136 columns took ~2 minutes, 3 seconds with out -F and less than a second with the -F option:

$ time grep -w -f shortlist.txt largefile.tsv > out_withoutF
 
real    2m3.181s
user    2m0.780s
sys     0m2.196s
$ time grep -wF -f shortlist.txt largefile.tsv > out_withF
 
real    0m0.568s
user    0m0.500s
sys     0m0.060s

Pairing a device with a Logitech unifying receiver in Linux

My girlfriend’s keyboard and mouse stopped working some time ago. It turned out that her Logitech unifying receiver (a small USB dongle for keyboard and mouse) was a bit broken, only when twisted in a certain way it would work. So, I called Logitech, explained the situation and they offered to send us a replacement for free. Well done Logitech support!

Now, since we both use Linux as our main OS, the question was how to pair the mouse and keyboard with the new receiver. Logitech provides a piece of Windows software, but nothing for Linux. It turns out it’s not that difficult and you can find various little C programmes that do it for you. I tried Travis Reeder’s solution and it worked like a charm on my Ubuntu 12.04 machine.

These are the steps I took.
First I switched off the keybord and the mouse, then ran the following:

$ git clone https://github.com/treeder/logitech_unifier.git
Cloning into 'logitech_unifier'...
remote: Counting objects: 35, done.
remote: Compressing objects: 100% (26/26), done.
remote: Total 35 (delta 11), reused 33 (delta 9)
Unpacking objects: 100% (35/35), done.
$ cd logitech_unifier/
$ ./autopair.sh 
Logitech Unified Reciever unify binary not compiled, attemping compilation
Logitech Unified Reciever unify binary was successfully compiled
Auto-discovering Logitech Unified Reciever
Logitech Unified Reciever found on /dev/hidraw0!
Turn off the device you wish to pair and then press enter
[sudo] password for lennart: 
The receiver is ready to pair a new device.
Switch your device on to pair it.

I ran the autopair.sh script twice, once for the mouse and once for the keyboard.

Thanks Travis!

Older posts

© 2018 Lennart's weblog

Theme by Anders NorenUp ↑