17. How to configure a new architecture

17. How to configure a new architecture
Prev		Next

The next steps are most easily done as user portbuild.

The following assumes you have already run mkportbuild.

For each branch that will be supported, do the following:
- Kick-start the build for the branch with
  # build create arch branch
If you are going to store your historical buildlogs and errorlogs on your head node's hard drive, you may skip this step. Otherwise:
Create an external directory and link to it:
Example 9. Creating and linking an external archive directory
# mkdir /dumpster/pointyhat/arch/archive # ln -s /dumpster/pointyhat/arch/archive archive

Note:
(Historical note that only applied to the original pointyhat.FreeBSD.org installation)
It is possible that /dumpster/pointyhat will not have enough space. In that case, create the archive directory as /dumpster/pointyhat/arch/archive and symlink to that.
Populate clients as usual.
Edit portbuild.conf from one of the ones for another architecture. addarch will have created a default one for you.
Create customized portbuild.machinename.conf files as appropriate.
If you need to create any tunnels:
1. Make a private configuration directory:
  # mkdir /a/portbuild/conf/arch
2. In that directory: create any dotunnel.* scripts needed.

As srcbuild:

Add arch to SUPPORTED_ARCHS in /a/portbuild/admin/conf/admin.conf.
Add the arch directory to /a/portbuild/admin/scripts/zbackup. (This is a hack and should go away.)

Enable the appropriate arch entry for /a/portbuild/scripts/dologs to the portbuild crontab. (This is a hack and should go away.)

Again as srcbuild:

For each branch that will be supported, do the following:
- Create bindist.tar.

Prev		Next
16. How to rebase on a supported FreeBSD branch	Home	18. How to configure a new head node (pointyhat instance)