Tag Info

Hot answers tagged

60

You need to specify the frontend as `noninteractive' and it will save your current settings. dpkg-reconfigure will take the current system settings as gospel, so simply change your timezone the way you would normally and run it with the non-interactive flag e.g. for me to change to "Europe/Dublin" where I am: # echo "Europe/Dublin" > /etc/timezone ...


36

Note: in the following commands, a command beginning with 'root#' means it needs to be run as root. To find which files were installed by a package, use dpkg -L: $ dpkg -L $package apt-file can tell you which files will be installed by a package before installing it: root# apt-get install apt-file root# apt-file update $ apt-file list $package Or if ...


27

I also posted this question on the puppet users group and this was a response that I got back. If you add ensure latest it will check the source file against the currently installed package and install the new one if it is latest. I'm still not sure how you would roll back to an older version, but this seems to solve my problem for now. package { ...


25

The apt history is in /var/log/apt/history.log as said in a comment above. That said, this will not list packages that were installed manually, using dpkg or GUIs such as gdebi. To see all the packages that went through dpkg, you can look at /var/log/dpkg.log.


15

Use the cruft package.


15

Use mk-build-deps which is part of devscripts.


15

I don't thinks so, in Ubuntu md5 checksums are only stored for certain files. For any given package the list of files that have checksums can be found in /var/lib/dpkg/info/<package>.md5sums e.g /var/lib/dpkg/info/openssh-server.md5sums These generally don't contain a complete list of the files that have been installed by a package e.g. ...


15

The answer was to set the perl:locale's as per: http://www.thomas-krenn.com/en/wiki/Perl_warning_Setting_locale_failed_in_Debian export LANGUAGE=en_US.UTF-8 export LANG=en_US.UTF-8 export LC_ALL=en_US.UTF-8 locale-gen en_US.UTF-8 dpkg-reconfigure locales This sovles the issue but does add unwanted time to the vagrant up provisioning.


13

Will remove a package and all configuration files. apt-get --purge remove package If it is still giving you issues, try reinstalling then purging: apt-get --reinstall install package apt-get --purge remove package You still might need to: killall package Reboot and if it started again run: update-rc.d remove package Story Time: Once for some ...


11

dpkg-statoverride is the management tool for a database maintained by dpkg that contains owner and mode settings for given file paths. Invoked as in your question, it will do two things: Set the owenrship/mode for the given files immediately if they exist (--update) and store a new entry in the override file (--add). The latter ensures that further dpkg ...


10

PGP-signing Debian packages is not necessary if you built them for your own private use. It's just the common authentication method for (official) Debian developers when they upload new packages into the Debian "unstable" branch. To avoid the error message just use: dpkg-buildpackage -uc -us (see also the manpage of dpkg-buildpackage)


9

As in dpkg/1.17.2, it implements --verify option, according to this debian bug report. Note this is a relatively new change to dpkg. Date: Thu, 05 Dec 2013 04:56:31 +0100 line in the dpkg v1.17.2 package shows this. Here is a brief description of --verify action quoted from the man page of dpkg. -V, --verify [package-name...] Verifies the ...


8

He fixed it reinstalling the files that appeared there. So you might want to try something like this: for package in $(apt-get upgrade 2>&1 | grep "warning: files list file" | sed "s/.*'//; s/://"); do apt-get install --reinstall "$package"; done Be aware, that running this command takes some time, as we cycle through every package.


8

You can print the control file and some other information with dpkg -I package.deb, or use dpkg -e package.deb to extract only the control information files. Also, you can do a dry run to see what dpkg would do with --dry-run: dpkg --dry-run -i package.deb


8

Okay, I figured out how to do this: aptitude search ~i -F "%s# %p" Which of course can easily be grepped to find items from the “universe” repository: aptitude search ~i -F "%s# %p" | grep universe


8

Install and use the debian package named equivs. ... Another use is to circumvent dependency checking: by letting dpkg think a particular package name and version is installed when it isn't man equivs-build


8

I'm not positive on the settings in Lenny but I know that in Squeeze, sudo is configured with env_reset meaning it will strip out all but a very few select env variables before running the command. This means the DEBIAN_FRONTEND variable you set is never actually making it to the apt-get install --yes --force-yes r-base. If you have full access to sudo, ...


7

On the reference installation (only once): dpkg-query -W -f='${Package}\n' | sort > baselist.txt (The following assumes bash) To get the packages added from the reference installation (this doesn't show what was removed): comm -1 -3 baselist.txt <(dpkg-query -W -f='${Package}\n' | sort) Even better, avoiding copy of baselist.txt: comm -1 -3 ...


7

This functionality is provided by rpm, not yum: rpm -ql [packagename] From the documentation: The general form of an rpm query command is rpm {-q | --query} [PACKAGE_NAME] [query-options] Information selection options: -l, --list List files in package.


6

rpm -ql packagename is roughly equivalent. You should think of yum as similar to apt-get and rpm as roughly equivalent to dpkg. yum deals with packages in terms of repositories, and rpm deals with individual packages. Ubuntu actually provides a cheat sheet on similar actions: Switching between RedHat and Ubuntu


6

I got the error message to go away by putting the following in my provisioning script, prior to any apt-get calls: export DEBIAN_FRONTEND=noninteractive This makes debconf use a frontend that expects no interactive input at all, preventing it from even trying to access stdin.


6

We see that df claims only 19MB is used in /boot and 151MB is available. But the directory listing shows far more than 19MB of files! Therefore I would guess that the /boot filesystem has been corrupted. Unmount it and check it: umount /boot fsck -f /boot


6

Above answer didn't worked for me completely. Couple of packages, namely libc6, were still showing this error. I found a solution on one forum. Bug is that new version of libc didn't echoed list of files to correct file. Instead of echoing it to /var/lib/dpkg/info/libc6:amd64.list it echoed it to /var/lib/dpkg/info/libc6.list It can be fixed by running ...


6

http://www.debianadmin.com/clone-your-ubuntu-installation.html and /var/adm/apt/history.log


5

When you make a different version of a package (as you did, since you changed the build rules), you should really add a changelog entry. That way, your package will have a different version number (so you'll know immediately that it's not the standard package from the distribution), and you'll have a trace of what you changed. Edit debian/changelog, and add ...


5

Actually I can use dpkg-checkbuilddeps which shows the build dependencies. That gets me 99% of what I need


5

A more concise variant of the debconf solution mentioned in another answer is the following: echo 'libc6 libraries/restart-without-asking boolean true' | sudo debconf-set-selections I just used this solution successfully when upgrading glibc on Debian 7 (wheezy). For the curious, the way I found the right configuration variable to set was as follows: ...


5

You should be able to regenerate the directory using something like this: for i in $(dpkg -l|awk '/^ii/ {print $2}') do apt-get --reinstall -y install $i done What that snippet should do is reinstall every package you already have installed. It will also upgrade everything. If you don't want to do that, you'll have to extend the script to install ...


5

Come up with a better user naming scheme... (or force "kdm" to use different login credentials) I've had to learn this lesson over the years as I inherited commercial Unix systems with three-letter usernames. Moving those servers to Linux exposed conflicts with system service accounts. The worst case was Randy P. McDonald, or userID "rpm". The RPM package ...


4

You probably want to take a look at all the documentation surrounding Debtags. It's... not as simple as you might think. No, I don't know why, either.



Only top voted, non community-wiki answers of a minimum length are eligible