I am working on an RPM which unpackages a tar file into an RPM whenever I run rpmbuild. I have two questions around the process:
Is my process of unpackaging the tar file into the RPM correct?
When I install the actual rpm onto a server, I'd like it to run a script inside the RPM which I have copied in called install.sh. How do I do that?
%build
# let's skip this for now
%install
mkdir -p %{buildroot}
chmod 755 ~/rpmbuild/SOURCES/bin/*
cp -frv ~/rpmbuild/SOURCES/bin/* %{buildroot}
%files
/*
%changelog
# let's skip this for now
Generally RPM support pre and post installation/uninstall scripts. And they are defined with %pre, %post, %preun and %postun. So if you are sure this file (install.sh) already exist you can run it on this way:
%pre
/path/to/install.sh
or
%post
/path/to/install.sh
I want to build a rpm package from a spec file(hello-world.spec).The command and error output are given below
Command1:
rpmbuild -ba hello-world.spec
ErrorOutput1:
Executing(%prep): /bin/sh -e /var/tmp/rpm-tmp.z4GoQn
+ umask 022
+ cd /root/rpmbuild/BUILD
+
: not foundm-tmp.z4GoQn: 28: /var/tmp/rpm-tmp.z4GoQn:
error: Bad exit status from /var/tmp/rpm-tmp.z4GoQn (%prep)
RPM build errors:
Bad exit status from /var/tmp/rpm-tmp.z4GoQn (%prep)
Content of my spec file is
Name: hello-world
Version: 1
Release: 1
Summary: Most simple RPM package
License: FIXME
%description
This is my first RPM package, which does nothing.
%prep
# we have no source, so nothing here
%build
cat > hello-world.sh <<EOF
#!/usr/bin/bash
echo Hello world
EOF
%install
mkdir -p %{buildroot}/usr/bin/
install -m 755 hello-world.sh %{buildroot}/usr/bin/hello-world.sh
%files
/usr/bin/hello-world.sh
%changelog
# let's skip this for now
My System/software details are
OS: Ubuntu 16.04
RPM version:4.12.0.1
The list of contents in rpmbuild folder is
Command2:
:~/rpmbuild# ls
Output2:
BUILD BUILDROOT RPMS SOURCES SPECS SRPMS
Can anybody help on this issue?
Check how your files is being generated. If you are passing files from windows to Linux, usually, there are different EOL for the files. We had a similar issue, I recommend you to change all the files to Unix format. I'm sharing with you how to reproduce your issue, if you edit that file in a Linux environment, like "nano", the problem disappears.
I'm new to the rpmbuild, googled and copy one spec file paramenters and created rpm successfully when i ran it the rpm nothing is creating
var.tar.gz -- this is my tar file. it has
var/
var/www/
var/www/html/index.html
My spec file
[root#kaka1 SPECS]# vi var.spec
Name: var
Version: 1
Release: 0
Summary: Xiph Streaming media server that supports multiple formats.
Group: Applications/Multimedia
License: GPL
URL: http################
Source: var-1.0.tar.gz
Prefix: %{_prefix}
Packager: Sukama
BuildRoot: %{_tmppath}/%{name}-root
%description
easily and supports open standards for commuincation and interaction.
%prep
rm -rf /root/rpmbuild/BUILD/*
%setup -n var
gzip -dc /root/rpmbuild/SOURCES/var-1.0.tar.gz| tar -xvvf -
if [ $? -ne 0 ]; then
exit $?
fi
could you help me where can i should mention my files to create under root?
First, don't use name like var, usr as package name, use that's more relevant to what the package may represent. Here is a simple spec file that should install index.html into the /var/www directory.
%define debug_package %{nil}
Summary: Simple SPEC
Name: simple
Version: 0
Release: 1
License: NONE # Use your suitable license
Group: NONE
URL: NONE
Source0: %{name}-%{version}.tar.gz
BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-root
## Following lines are commented out, but that's the way to go
#BuildRequires: httpd # This needs https RPM during your build
#Requires: httpd # This would require httpd to be installed for this RPM
%description
%prep
%setup -q # This untars the source
%build # currently empty
%install # This moves into the source directory "simple-0" (see below)
rm -rf $RPM_BUILD_ROOT
mkdir -p $RPM_BUILD_ROOT/var/www
install -m 0644 index.html %{buildroot}/var/www/index.html
%clean
rm -rf $RPM_BUILD_ROOT
%post
%files
%defattr(-,root,root,-)
%doc
/var/www/index.html # Your packages owns the file, but not the directories
%changelog
* Web Feb 14 2018 <user#localhost> - 0.0
- Initial build.
For the above .spec file, one would need a tarball named simple-0.tar.gz. This maps to the Source0 name that is used inside the .spec. The content in this case would be a directory and a file: simple-0/index.html.
I have a problem with running preinstall script within PKG I created for Mac OS X on linux machine.
My problem is, that preinstall script is not executed and installer finishes with sucess, even my script should cause the failure.
I have prepared following directory structure on the hard-drive:
./Applications
./Resources
./Scripts
by running of following commands:
mkdir -p Payload.pkg
chmod -R a+w Applications
chmod a+x Scripts/preinstall
mkbom -u 0 -g 0 Applications Payload.pkg/Bom
find Applications | cpio --owner 0:0 -H odc -o | gzip -c > Payload.pkg/Payload
pushd Scripts
find . | cpio --owner 0:0 -H odc -o | gzip -c > Payload.pkg/Scripts
popd
chmod a+x Payload.pkg/Scripts
xar --compress none -cf MyApp.pkg Distribution Resources Payload.pkg
I have created package with following structure:
My package.pkg
+--Distribution
+--Resources
+--welcome.rtf
+--Payload.pkg
+--Scripts
+--Bom
+--PackageInfo
+--Payload
my preinstall scripts has following content:
#!/bin/bash
##preinstall
pathToScript=$0
pathToPackage=$1
TargetLocation=$2
targetVolume=$3
exit 1
Which should be causing failure of the installer. However it does not. It seems, that script is not invoked.
Could somebody help me please?
If you want a script to be called, it needs to be included in the <scripts> section of PackageInfo like so:
<scripts>
<preinstall file="./preinstall"/>
</scripts>
Source: https://hogliux.github.io/bomutils/tutorial.html
My production server (Centos 5.9) won't compile nodejs, possibly because it's gcc is only 4.1.2 (4.2 or above is recommended) so I've trying to install the binaries.
$ wget http://nodejs.org/dist/v0.10.22/node-v0.10.22-linux-x86.tar.gz
$ tar -zxvf node-v0.10.22-linux-x86.tar.gz
$ cd node-v0.10.22-linux-x86
$ sudo cp bin/* /usr/local/bin
$ sudo cp -R lib/* /usr/local/lib
$ sudo cp -R share/* /usr/local/share
And now for testing:
$ node -v # => v0.10.22
$ man node # looks fine
$ npm -v # UH OH, PROBLEM - Cannot find module 'npmlog'
Now (keeping in mind I'm a complete beginner at node) I did some searching and found there's an environment variable called NODE_PATH, so I tried:
$ export NODE_PATH=/usr/local/lib/node_modules
$ npm -v # SAME PROBLEM - Cannot find module 'npmlog'
So then I found out where npmlog lives and tried modifying NODE_PATH accordingly:
$ find /usr/local/lib -name npmlog # => /usr/local/lib/node_modules/npm/node_modules/npmlog
$ export NODE_PATH=/usr/local/lib/node_modules/npm/node_modules
$ npm -v # DIFFERENT PROBLEM - Can't find '../lib/npm.js'
At this stage, after more unhelpful googling, I decided I was in over my depth and decided to ask for help. Can anyone tell me what I'm doing wrong?
It is much faster to do clean NPM reinstall which will remove "broken" links:
wget https://npmjs.org/install.sh
chmod +x install.sh
sudo ./install.sh
Then it will ask you to remove old NPM link
Using Node Version Manager
Use a Node version manager like nvm to handle installation and version management for you. After you install nvm you can simply install any Node version, for example nvm install 8.
But if you just want to install the binary yourself, see below:
Using apt-get
In special cases where you need a system wide Node installation, you can use apt-get:
curl -sL https://deb.nodesource.com/setup_8.x | sudo -E bash -
sudo apt-get install -y nodejs
The above snippet will install the latest Node 8.
Installing the Binary Manually
In order to install the binary manually, all you need to do is to download the binary and create a bunch of symbolic links. Execute the commands below one after the other, and it should do the job. I have also written a shell script that does it for you if that is easier (see the bottom of the answer). Hope that helps.
Make sure to use the correct download link for your OS architecture (i.e. either 32-bit or 64-bit) for wget on the second line.
ME=$(whoami) ; sudo chown -R $ME /usr/local && cd /usr/local/bin #adding yourself to the group to access /usr/local/bin
mkdir _node && cd $_ && wget https://nodejs.org/dist/v8.11.4/node-v8.11.4-linux-x64.tar.xz -O - | tar zxf - --strip-components=1
ln -s "/usr/local/bin/_node/bin/node" .. # Making the symbolic link to node
ln -s "/usr/local/bin/_node/lib/node_modules/npm/bin/npm-cli.js" ../npm ## making the symbolic link to npm
Here is a shell script that downloads and installs all the components. If you use this script to install Node, you can use the uninstall script to uninstall it.
Installing Node
#! /bin/bash
# run it by: bash install-node.sh
read -p " which version of Node do you need to install: for example 8.11.4 (or any other valid version): " VERSIONNAME
read -p " Are you using a 32-bit or 64-bit operating system ? Enter 64 or 32: " ARCHVALUE
if [[ $ARCHVALUE = 32 ]]
then
printf "user put in 32 \n"
ARCHVALUE=86
URL=https://nodejs.org/dist/v${VERSIONNAME}/node-v${VERSIONNAME}-linux-x${ARCHVALUE}.tar.gz
elif [[ $ARCHVALUE = 64 ]]
then
printf "user put in 64 \n"
ARCHVALUE=64
URL=https://nodejs.org/dist/v${VERSIONNAME}/node-v${VERSIONNAME}-linux-x${ARCHVALUE}.tar.gz
else
printf "invalid input expted either 32 or 64 as input, quitting ... \n"
exit
fi
# setting up the folders and the the symbolic links
printf $URL"\n"
ME=$(whoami) ; sudo chown -R $ME /usr/local && cd /usr/local/bin #adding yourself to the group to access /usr/local/bin
mkdir _node && cd $_ && wget $URL -O - | tar zxf - --strip-components=1 # downloads and unzips the content to _node
cp -r ./lib/node_modules/ /usr/local/lib/ # copy the node modules folder to the /lib/ folder
cp -r ./include/node /usr/local/include/ # copy the /include/node folder to /usr/local/include folder
mkdir /usr/local/man/man1 # create the man folder
cp ./share/man/man1/node.1 /usr/local/man/man1/ # copy the man file
cp bin/node /usr/local/bin/ # copy node to the bin folder
ln -s "/usr/local/lib/node_modules/npm/bin/npm-cli.js" ../npm ## making the symbolic link to npm
# print the version of node and npm
node -v
npm -v
Uninstalling Node
#! /bin/bash
# run it by: ./uninstall-node.sh
sudo rm -rf /usr/local/bin/npm
sudo rm -rf /usr/local/bin/node
sudo rm -rf /usr/local/lib/node_modules/
sudo rm -rf /usr/local/include/node/
sudo rm -rf /usr/local/share/man/man1/node.1
sudo rm -rf /usr/local/bin/_node/
I had a problem like that, but with iojs. However it should be the same procedure:
(Assuming that you've got a file matching node-v*-linux-x64.tar.gz in your current directory):
# In case of iojs you need to replace the occurrences of 'node' with 'iojs'
# Extract the downloaded archive with the linux-x64 version of node
tar zxf node-v*-linux-x64.tar.gz
# Move the extracted folder (./node-v*-linux-x64/) to /opt/
mv ./node-v*-linux-x64/ /opt/
To make the binary files available in your shell, create some softlinks inside the /usr/bin/ directory:
# Create a softlink to node in /usr/bin/
ln -s /opt/node-v*-linux-x64/bin/node /usr/bin/node
# Create a softlink to npm in /usr/bin/
ln -s /opt/node-v*-linux-x64/bin/npm /usr/bin/npm
# Create a softlink to iojs in /usr/bin (this step can be omitted if you're using node)
ln -s /opt/node-v*-linux-x64/bin/iojs /usr/bin/iojs
Notice: If you'd like to access the cli of some globally installed node modules (for example bower, typescript or coffee-script), you're required to create a softlink to each of those executables in the /usr/bin/ directory.
Alternatively you could just add the bin directory of your node installation directory (e.g. /opt/node-v*-linux-x64/) to the PATH environment variable: (you should use the absolute path for this!)
# create a new .sh script in /etc/profile.d which adds the directory to PATH
echo "export PATH=$PATH:/opt/node-v0.12.3-linux-x64/bin" > /etc/profile.d/node-npm.sh
This change will take effect after logging out and in again.
Both methods worked for me (I use a linux desktop version of Ubuntu 14.04/15.04 with GNOME 3).
I had the same issue reported here. Fixed it by removing /usr/local/bin/npm and replacing it with a symlink to /usr/local/lib/node_modules/npm/bin/npm-cli.js
$ ls -l /usr/local/bin/
node
npm -> /usr/local/lib/node_modules/npm/bin/npm-cli.js
$ npm -v
1.3.17
wget <node archive url from nodejs.org>
cd /usr/local
sudo tar --strip-components 1 -xf <path to node archive>
You can run node and npm right away.
It used to be documented in the README inside the archive in older versions.
I had the same problem and I was able to resolve it by creating symlinks instead of copying the binaries.
$ cd /usr/local/src
$ wget http://nodejs.org/dist/v0.10.24/node-v0.10.24-linux-x64.tar.gz
$ tar -zxvf node-v0.10.24-linux-x64.tar.gz
$ cd node-v0.10.24-linux-x64
$ sudo cp -R lib/* /usr/local/lib
$ sudo cp -R share/* /usr/local/share
$ ln -s /usr/local/src/node-v0.10.24-linux-x64/bin/node /usr/local/bin/node
$ ln -s /usr/local/src/node-v0.10.24-linux-x64/bin/npm /usr/local/bin/npm
$ node -v
v0.10.24
$ npm -v
1.3.21
I tend to use nave to install the binaries. Use wget to download the nave.sh file and then us it to install node. Nave is also nice to have around in case one of your production apps requires a different version of node than what's installed globally.
$ wget https://raw.github.com/isaacs/nave/master/nave.sh
$ sudo bash nave.sh usemain 0.10.22
You can use GNU stow to make symbolic links of those binaries in /usr/local properly with one command. Stow also allows you to easily remove Node js from /usr/local at a later time and swap multiple versions of Node js.
$ # first, install stow
$ mkdir /usr/local/stow # if it doesn't exist
$ # then, place software binary package in /usr/local/stow
$ cd /usr/local/stow
$ stow <package_name> # install / add sym links
$ source $HOME/.bash_profile # reload your environment
$ # node -v and npm -v should now work
$ stow -D <package_name> # uninstall / remove sym links
These steps worked for me with node-v0.10.17-linux-x64.
In the man page of cp in Mac OS X:
Symbolic links are always followed unless the -R flag is set, in which case symbolic links are not followed, by default.
When you execute sudo cp bin/* /usr/local/bin, the symbolic link bin/npm is followed.
Actually, bin/npm is linked to ../lib/node_modules/npm/bin/npm-cli.js, so cp will copy npm-cli.js to /usr/local/bin. That's why you get an error.
I had the same problem.
The problem is the npm excutable in /usr/local/bin.
The way I solved it was:
sudo rm /usr/local/bin/npm
sudo ln -s "/usr/local/lib/node_modules/npm/bin/npm-cli.js" /usr/local/bin/npm
In Ubuntu there is a .bashrc file which sets path to binaries.
By default, there is path set for bin in home directory. Perhaps you can create bin directory in your home directory and move the binaries there. Reboot your system and try executing the command node
I faced the same problem. So, I symlinked node and npm from ./bin/ to /usr/local/bin
If someone is interested in using Docker, in the Dockerfile,
ENV NODE_VERSION 8.10.0
RUN wget https://nodejs.org/dist/v$NODE_VERSION/node-v$NODE_VERSION-linux-x64.tar.xz
RUN tar -xJvf node-v$NODE_VERSION-linux-x64.tar.xz -C /usr/local/
ENV NODEJS_HOME /usr/local/node-v$NODE_VERSION-linux-x64
ENV PATH $NODEJS_HOME/bin:$PATH
RUN node --version
RUN npm --version