Version 5 (modified by Bruno Cornec, 5 years ago) (diff)


Setup and Configuration


In order to have a better understanding of setup and capabilities, have a look at the Lab documentation.

Configuration file parameters

Look at the documentation in progress at

Dealing with Virtual Machines and Virtual Environments

Virtual Machines

In order to use a new VM (Virtual Machine) with pb, you have to follow a 4-step approach:

  1. Create an appropriate VM using pb (or not if you prefer to create it outside with your preferred tools)
    pb -p prj -m os-ver-arch -i /path/to/os.ver.arch.iso newver

where you pass to the pb command the ISO image of your distribution media. If you prefer to deploy using PXE e.g. you have to use instead the qemu command directly. At the end of the installation, pb makes recommendations on what minimum packages have to be present (sshd and sudo e.g.) and their minimal setup. Something that has to be done manually, at worst, before being able to launch the second phase.

  1. Set up the VM so it gets all what is needed to run pb in it:
    pb -p prj -m os-ver-arch setupvm

This step will install pb inside your VM, create a dedicated account for building packages, exchange dedicated ssh keys from your own account, and setup sudo rights so the build can be made non-interactively. The first time ou call that command, it will ask you the root passwd of the VM, in order to setup up auhtentication. Each time you want to re-align pb versiosn between your host and the guest, you'll have to re-run that command. The root passwd won't be asked this time, so it can be automated easily. (If you don't precise a specific VM to work on, all the the one declared in your configuration file will be updated).

  1. Prepare the VM for your project
    pb -p prj -m os-ver-arch prepvm

This step will install all the required dependencies for your prj project inside the VM. It will analyze the requirements coming from the spec file, or control file and launch the installation of the required packages.

  1. Finally you're able to use your VM to build your software packages:
    pb -p prj -m os-ver-arch sbx2vm

Note that you can specify multiple VMs separated by ','.

Virtual Environments

Also note that you can do the same with Virtual Environments (VE) replacing the vm suffix by the ve one. This will work for Docker containers as well, and you can use the -T docker option for that.

Gentoo Container


$ docker run -ti gentoo/stage3-amd64 /bin/bash
# emerge-webrsync
# exit
$ docker ps -l
CONTAINER ID    IMAGE                 COMMAND      CREATED             STATUS                    PORTS     NAMES
a40e2247b233    gentoo/stage3-amd64   "/bin/bash"  9 minutes ago       Exited (0) 3 seconds ago            sick_hodgkin
$ docker commit a40e2247b233 gentoo:latest
$ pb -p prj -m gentoo-nover-x86_64 -T docker -i gentoo newve

Non upstream usage of

Using patches

If you're not upstream, you may need to apply patches to the software you're packaging. In short, if you create a pbpatch directory (ex: under your package directory, and order your patches in it (00-patch-to-do-a, 01-patch-to-do-b, ...) , and then use the macros PBPATCHSRC and PBPATCHCMD in your spec file (ex: you should be able to produce rpms the way you want.