You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Use a single partition for the Vagrant images (#115)
* Use a single partition for Vagrant images
There is no performance penalty for using a swapfile instead of a swap
partition; by not defining a swap of boot partition, we allow our users
to create a swapfile themselves if they need one.
* Use Kickstart to set the bootloader timeout
* Remove several unneccessary packages
- plymouth displays a bootup screen, not needed by Vagrant
- manpages don't bring anything if we use --excludedocs
- linux-firmware isn't useful for VM hardware
- support for ext[234] and btrfs isn't needed, we use XFS
* Add the man pages again to the image
* Configure swapping to a file
We use a 2GiB file as swap, as recommended by Red Hat for systems with
up to 4G RAM.
* Add cifs-utils to the images (for Windows users)
* Switch the filesystem from xfs to ext4
All CentOS Linux 7 images in the last few months are unable to boot due
to apparent XFS filesystem corruption. This is 100% reproducible with
both Image Factory and Packer when using libvirt (the images for
VirtualBox are working). It might be caused by qemu-img, since the
initial boot performed by Packer immediately after the installation is
successful.
0 commit comments