Linux Central

Restore a drive/partition from backup partition table entries

/n/ — Wed, 29 Jul 2009 11:51:13 +0000

Having detailed knowledge of your partition table is worth it’s weight in gold if you’re ever faced with a disk that just lost its partition table entries.

I was recently faced with such a tragedy, and I was only saved by virtue of the fact that I had the partition table entries to hand. I had just deleted an ext3 partition on my external drive using the Gnome Partition Editor, and then queued up two operations:

Create a fat32 file system in its place
Assign it a label

When I applied these operations – BOOM! – the whole partition table on the external drive just disappeared. This was the drive that held my personal data, my photos and my entire media collection – basically any data of value to me. Thankfully, I had a backup of my partition table entries and instructions on how to recreate the partition table using these entries.

So, remember to back up your partition table entries to a file and store it offsite or at multiple locations on your home network. These days there are plenty of free offsite backup options like Dropbox and Ubuntu One, so there is never an excuse not to.

Backing up your partition table

We’ll use the gpart utility to get a snapshot of the partition table, so install the package of the same name if you don’t find the gpart command on your system.


sudo apt-get install gpart

Next, run it against the disk who’s partition table needs to be backed up. Remember to do this against every disk you have, and to do this every time you change the partition table on a disk.

If you’re not sure which device corresponds to which drive, the mount command is your friend:


$ mount
/dev/sda5 on / type ext3 (rw,relatime,errors=remount-ro)
tmpfs on /lib/init/rw type tmpfs (rw,nosuid,mode=0755)
proc on /proc type proc (rw,noexec,nosuid,nodev)
sysfs on /sys type sysfs (rw,noexec,nosuid,nodev)
varrun on /var/run type tmpfs (rw,nosuid,mode=0755)
varlock on /var/lock type tmpfs (rw,noexec,nosuid,nodev,mode=1777)
udev on /dev type tmpfs (rw,mode=0755)
tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev)
devpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=620)
fusectl on /sys/fs/fuse/connections type fusectl (rw)
lrm on /lib/modules/2.6.28-14-generic/volatile type tmpfs (rw,mode=755)
/dev/sda2 on /media/sda2 type fuseblk (rw,nosuid,nodev,allow_other,blksize=4096)
securityfs on /sys/kernel/security type securityfs (rw)
binfmt_misc on /proc/sys/fs/binfmt_misc type binfmt_misc (rw,noexec,nosuid,nodev)
//192.168.1.77/HDD_1_1_1 on /media/Elements1 type cifs (rw,mand,noexec,nosuid,nodev)
//192.168.1.77/HDD_1_1_2 on /media/Elements2 type cifs (rw,mand,noexec,nosuid,nodev)
192.168.1.77:/share/hdd/data/HDD_1_1_1 on /media/HDD_1_1_1 type nfs (rw,users,noexec,nosuid,nodev,intr,addr=192.168.1.77)
192.168.1.77:/share/hdd/data/HDD_1_1_2 on /media/HDD_1_1_2 type nfs (rw,users,noexec,nosuid,nodev,intr,addr=192.168.1.77)
192.168.1.77:/share/hdd/data/HDD_1_1_3 on /media/HDD_1_1_3 type nfs (rw,users,noexec,nosuid,nodev,intr,addr=192.168.1.77)

Once you’ve identified a disk, e.g /dev/sdb, it’s time to run the gpart command to reveal it’s partition table:


$ sudo gpart /dev/sdb

Begin scan...
Possible partition(Linux ext2), size(410002mb), offset(0mb)
Possible partition(DOS FAT), size(410002mb), offset(410002mb)
Possible partition(Linux ext2), size(133861mb), offset(820005mb)
End scan.

Checking partitions...
Partition(Linux ext2 filesystem): primary
Partition(DOS or Windows 95 with 32 bit FAT, LBA): primary
Partition(Linux ext2 filesystem): primary
Ok.

Guessed primary partition table:
Primary partition(1)
 type: 131(0x83)(Linux ext2 filesystem)
 size: 410002mb #s(839685352) s(63-839685414)
 chs:  (0/1/1)-(1023/254/63)d (0/1/1)-(52267/254/58)r

Primary partition(2)
 type: 012(0x0C)(DOS or Windows 95 with 32 bit FAT, LBA)
 size: 410002mb #s(839685420) s(839685420-1679370839)
 chs:  (1023/254/63)-(1023/254/63)d (52268/0/1)-(104535/254/63)r

Primary partition(3)
 type: 131(0x83)(Linux ext2 filesystem)
 size: 133861mb #s(274149224) s(1679370840-1953520063)
 chs:  (1023/254/63)-(1023/254/63)d (104536/0/1)-(121600/254/62)r

Primary partition(4)
 type: 000(0x00)(unused)
 size: 0mb #s(0) s(0-0)
 chs:  (0/0/0)-(0/0/0)d (0/0/0)-(0/0/0)r

Backup this information to a file and store in a secure location.

Restoring a drive from partition table entries

Next we want to restore the partition table from the entries we backed up earlier. For this we need the parted command. If your Linux machine is unusable (because it was on the disk you wish to recover), consider creating an Ubuntu LiveCD using someone else’s computer and running that on your machine to get to a terminal where you can run parted. You do not need Ubuntu as such; all we need is a distro with the ability to run the parted command.

Once you are on a terminal:

Run the command sudo parted /dev/sdb. It’ll drop you into the parted command shell, on a new line with the (parted) prompt
Enter “unit s” on the prompt to tell parted to use sectors as the unit to configure the partition table.
Enter “print” to check the disk details and confirm that you have the correct disk.
Next, type “rescue“. It should give you a Start? prompt. Enter the starting sector of the first partition from the partition table backup file. It next prompts you for the ending sector: End?. Provide the ending sector for the partition. Referring to the example below, the first partition ranges from sector 63 to 839685414 “s(63-839685414)” as evidenced by this snippet from earlier:

Primary partition(1)
 type: 131(0x83)(Linux ext2 filesystem)
 size: 410002mb #s(839685352) s(63-839685414)
 chs:  (0/1/1)-(1023/254/63)d (0/1/1)-(52267/254/58)r

If all goes well, it should respond by saying something like “A ext3 primary partition was found at 63s -> 839685414s. Do you want to add it to the partition table? Yes/No/Cancel?“. Say Yes and proceed.
Follow these steps for each partition you had on the disk.
Once done, type “quit” to exit the parted command.

For more detailed instructions on recovering lost partitions, look at the Ubuntu Data Recovery Community Guide.

Provided below is a sample output of a recovery session:


$ sudo parted /dev/sdb
GNU Parted 1.8.8
Using /dev/sdb
Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) unit s                                                           
(parted) print                                                            
Model: WD 10EAVS External (scsi)
Disk /dev/sdb: 1953525168s
Sector size (logical/physical): 512B/512B
Partition Table: msdos

Number  Start  End  Size  Type  File system  Flags

(parted) rescue                                                           
Start? 63                                                                 
End? 839685414                                                            
Information: A ext3 primary partition was found at 63s -> 839685414s.  Do you want to add it to the partition table?
Yes/No/Cancel? yes                                                        
(parted) rescue                                                           
Start? 839685420                                                          
End? 1679370839                                                           
Information: A fat32 primary partition was found at 839685420s -> 1679370839s.  Do you want to add it to the partition
table?
Yes/No/Cancel? yes                                                        
(parted) rescue                                                           
Start? 1679370840                                                         
End? 1953520063                                                           
Information: A ext2 primary partition was found at 1679370840s -> 1953520063s.  Do you want to add it to the partition
table?
Yes/No/Cancel? yes                                                        
(parted) quit                                                             
Information: You may need to update /etc/fstab.

Tagged: data recovery, gpart, parted, partition table, restore

Improve responsiveness of Linux desktop: decrease “swappiness”

/n/ — Sat, 25 Jul 2009 16:54:45 +0000

The swappiness parameter controls how aggressively the kernel swaps processes out of physical memory and onto the swap disk.

When swappiness=0, the server kernel avoids swapping programs out of physical memory for as long as possible. When swappiness=100, the server aggressively swaps processes out of physical memory.

The Ubuntu default is 60 at time of writing. You can check this through the following command:

cat /proc/sys/vm/swappiness

We want to change this value, and lower it to (typically) 20, as swap space is much slower than physical memory and involves expensive I/O writes. Also, the constant swapping heats up the hard drive and lowers its life expectancy.

You can change this value temporarily ( till the next reboot ) through the following command:

sudo sysctl vm.swappiness=20

To make this change permanent, we’ll need to edit /etc/sysctl.conf

gksudo  gedit /etc/sysctl.conf

and add or modify this parameter.

vm.swappiness=20

If this parameter does not exist in the file, add it at the end. You’ll need to reboot to have this change take effect. Authoritative source for this tip is the Ubuntu Community Documentation Swap FAQ.

Open a terminal, and run the free command to check how much swap space is being used after implementing this tweak.

Sample output on my system after the tweak:


$free             total       used       free     shared    buffers     cached
Mem:       1017416    1001440      15976          0      14304     233252
-/+ buffers/cache:     753884     263532
Swap:      1004020     133460     870560

Tagged: Performance, swappiness, sysctl

Purge packages that are deinstalled

/n/ — Sat, 25 Jul 2009 12:13:47 +0000

Before starting on the command, an explanation of why this is generally useful. When a package is deinstalled (because it’s not required or because a newer version of the package is installed), the package is removed from the system, but the configuration files remain. To remove the config files too, you need to purge the package.

If you have manually configured a package earlier, and would like to retain the configuration files for when you’ll install the package in future, this is not a command you should run.

In all other cases, you should purge the packages in the interests of good housekeeping. If that isn’t enough motivation, it also means that future installs of the package get a blank state, and there are no potentially messy interactions of earlier version config files with newer packages.

With that out of the way, lets get down to bass tacks:

dpkg --get-selections | grep "[[:space:]]deinstall$" | awk '{print $1}' | sudo xargs dpkg -P

The first command in the pipeline returns the list of packages in the system, the second filters out the deinstalled packages, the third filters out the package names from the previous output, and the fourth purges each package. Consequently, only the fourth command requires sudo permissions.

There’s a downside to running this command – you cannot stop the command by hitting on the terminal as the xargs command spawns an individual process for each dpkg -P

On linux systems, there are typically many ways to achieve the same result, so here’s another way to purge deinstalled packages:

sudo aptitude purge $(aptitude search ~c -F "%p")

So you know what’s happening, here’s a brief explanation of what the command does. aptitude search -c returns the list of deinstalled packages. The -F “%p” argument ensures only the package names are printed. The outer aptitude purge command operates on the returned list of packages from the inner command.

An example of why you’d run this is to purge old linux kernels.

Before running such commands, it is imperative you know what you are running, and what the results are, so I’d suggest you run it in smaller chunks and examine the results before running the actual command.

Tagged: aptitude, cmd-fu, dpkg, pkg-management

Alternatives to Nautilus on a Gnome desktop

/n/ — Sat, 25 Jul 2009 11:41:34 +0000

Nautilus is the default file manager on the Gnome desktop. It has improved in leaps and bounds over the years, yet there are certain areas where more work is needed.

After one too many crashes trying to drag-and-drop folders, I had the motivation I needed to sit down to search and evaluate alternatives. I didn’t find any truly compelling alternatives, but a number of equally flawed replacements.

Thunar

Thunar is the file manager for the Xfce desktop environment. It has the look and feel of a Gnome application, and is a worthy replacement for Nautilus.


sudo apt-get install thunar

Dolphin

Dolphin is the KDE file manager. As expected, it looks a bit alien in a Gnome environment. However, if you’re willing to overlook this, it can be a great replacement for Nautilus. Personally, I find it more usable than Nautilus and the best replacement among the lot in this list. Dolphin does not pull in a ton of dependencies, but if you want to keep your system clean of KDE libraries then obviously this option is not for you.


sudo apt-get install dolphin

PCMan

A stand-alone file manager. Lightweight, fast. Not much to say except to try it out and see if you like it.


sudo apt-get install pcmanfm

ROX-filer

The file manager for the ROX desktop environment. This probably is the fastest and lightest replacement you can find; if speed is your primary concern, this is the alternative for you. It didn’t appeal to me. I found it’s behaviour of resizing the window to fit the contents of the folder a bit disconcerting. Your mileage may vary. Again, the best option is to try it and discard if you don’t like it. If you are so inclined, you can also integrate Rox-filer in place of Nautilus in Gnome.


sudo apt-get install rox-filer

Two-paned File Managers

Two-paned file managers are not really my cup of tea, but if you’re a fan of Midnight Commander and the like, you might consider these options too:

Krusader

The KDE version of the two-paned file manager. Obviously it does not look native on a Gnome desktop, but is quite surprisingly usable.


sudo apt-get install krusader

Gnome Commander

The Gnome version. Does not give the impression of a finished product.


sudo apt-get install gnome-commander

I tried out different alternatives each for limited periods, and at the end uninstalled the packages I don’t intend to use.

Barring obvious functionality and usability concerns, file manager preferences are dictated by quite personal tastes and what works for you will not be obvious till you actually use the software for a while. Also, your desktop environment dictates an application’s look and feel, and it may look significantly different from what you see in screen-shots on other sites.

For these reasons, I’d recommend you try out the application yourself.

Once you’ve tried and discarded an option, do not forget to uninstall the package completely.


sudo apt-get --purge remove pkg_to_uninstall

—

Versions: Nautilus 2.26.2

Tagged: desktop, file manager, Gnome, Nautilus

Clean out .deb package cache

/n/ — Fri, 24 Jul 2009 16:48:16 +0000

The debian packages are typically in /var/cache/apt/archives, ready to be used for a reinstallation.

If you are short on space, you can safely delete these:

sudo apt-get clean

To delete only unused .deb package files (deinstalled, older versions etc) use:

sudo apt-get autoclean

To delete orphaned packages (which aren’t required or used), use:

[sourecode language="shell"]
sudo apt-get remove --purge `deborphan`
[/sourcecode]
deborphan is a distinct command that lists the orphaned packages.

Tagged: apt-get, cmd-fu, deborphan, pkg-management

Turn system beep on/off

/n/ — Thu, 16 Jul 2009 11:08:36 +0000

Edit /etc/modprobe.d/blacklist and add these lines:


# get rid of system beep

blacklist pcspkr

Reboot to activate this line.To re-enable the speaker (that emits the beep), remove these lines and reboot.

Alternatively, disable it for the current session:


sudo rmmod pcspkr

To re-enable the system speaker for the current session:


sudo modprobe pcspkr

Tagged: cmd-fu, hacks, pcspkr

Search, generate or restore from list of installed packages

/n/ — Wed, 15 Jul 2009 12:49:19 +0000

Listing packages

To get the list of packages that start with pkg_name_begin:


dpkg --get-selections pkg_name_begin*

For a more generalized search, e.g where pkg_name_partial is text within the package name:


dpkg --get-selections | grep "[[:space:]]install$" | grep -i pkg_name_partial

To get a list of all installed packages on your debian system:


dpkg --get-selections | grep "[[:space:]]install$" | \awk '{print $1}'

It makes terrific sense to periodically generate and store this list for any eventually where you need to reinstall the system. You should even consider turning this into a script, and place into a weekly or monthly cron job that mails you the list.


dpkg --get-selections | grep "[[:space:]]install$" | awk '{print $1}' | tee installed_pkgs | mailx -s "List of installed packages" you@mailhost.com

Reinstalling packages

To install the packages from the list created earlier on top of a fresh Ubuntu install, follow these steps:


sudo apt-get upgrade

sudo apt-get dist-upgrade
dpkg --set-selections < installed_pkgs

sudo apt-get -u dselect-upgrade

Formatting woes: The third statement should read dpkg –set-selections < installed_pkgs, but wordpress won’t let it be.

Tagged: cmd-fu, dpkg, pkg-management

Check which package a file came from

/n/ — Wed, 15 Jul 2009 10:09:41 +0000


dpkg -S filename

This command returns the package(s) that install the given file. Useful when you want to retrieve the original file and/or find what it’s used for.

Tagged: cmd-fu, dpkg, pkg-management

Monitor CPU and hard disk temperature levels

/n/ — Tue, 14 Jul 2009 16:41:17 +0000

lm-sensors

Install the lm-sensors package: sudo apt-get install lm-sensors

Run command: sensors-detect

It detects the sensors on board your computer, and will ask you to confirm the ones it finds. It is generally safe to accept its recommendations. At the end, assuming it found a sensor, it will ask you if you want to automatically have them added to /etc/modules. If you say YES, these kernel modules will get loaded at next reboot. They let you query the on-board sensors through the sensor command.

$ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1:       +66.5°C  (crit = +105.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0:      +63.0°C  (crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1:      +66.0°C  (crit = +100.0°C)

You can also look this information up via GNOME sensors-applet applet (ksensors for KDE). If missing, install via sudo apt-get install sensors-applet.

Hddtemp

This command gives you the temperature of the hard disk given as a parameter.

$ sudo hddtemp /dev/sda1
/dev/sda1: Hitachi HTS541060G9SA00: 45°C

Alternatively, you can use the following command to get temperatures of all hard disks on the system:

$ nc localhost 7634
|/dev/sg0|Hitachi HTS541060G9SA00|45|C||/dev/sda|Hitachi HTS541060G9SA00|45|C|
$ echo "$(nc localhost 7634 | cut -d'|' -f4)"
45
$

Tagged: cmd-fu, lm-sensors, maintenance, monitoring

Force removal of a debian package

/n/ — Tue, 14 Jul 2009 14:12:38 +0000

dpkg -r --force-all package_name

DISCLAIMER: Highly unsafe command. Do not attempt to use without adequate knowledge or precaution; you could end up with a broken installation. A typical use of the command would be to reinstall a package without having to reinstall all other packages that depend on it.

Sometimes, you need to more than just remove the package; you need to purge it. Purging a package removes the package and it’s configuration files. That is achieved through this command:


dpkg --purge --force-all package_name


Versions: dpkg 1.14.24ubuntu1 (i386)

Tagged: cmd-line, dpkg, pkg-management