This repo is just a collection of data files for testing houndigrade.
| filename | partition | RHEL? | description |
|---|---|---|---|
| disks/centos_release | 1 | no | has CentOS release file |
| disks/misc_bad_files | 1 | no | has a "bad" non-text release file, yum.conf, RPM DB, and syspurpose file |
| disks/misc_empty | 1 | no | empty ext4 filesystem |
| disks/misc_ntfs | 1 | no | Windows-like NTFS filesystem |
| disks/rhcos_release | 1 | yes | has RHEL release file, ostree deployment |
| disks/rhel_all | 1 | yes | has RHEL product cert, release file, yum repo, signed RPM, and syspurpose file |
| disks/rhel_bad_syspurpose | 1 | yes | has RHEL release file and a corrupt 2K syspurpose file |
| disks/rhel_dnf | 1 | yes | has RHEL dnf repo enabled |
| disks/rhel_cert | 1 | yes | has RHEL product cert in /etc/pki/product/ |
| disks/rhel_cert2 | 1 | yes | has RHEL product cert in /etc/pki/product-default/ |
| disks/rhel_lvm | lvm | yes | has RHEL release file |
| disks/rhel_partitioned | 1 | no | is swap |
| disks/rhel_partitioned | 2 | yes | has RHEL release file |
| disks/rhel_partitioned | 3 | no | empty filesystem |
| disks/rhel_release | 1 | yes | has RHEL release file |
| disks/rhel_repo | 1 | yes | has RHEL yum repo enabled |
| disks/rhel_rpms | 1 | yes | has a Red Hat-signed RPM installed (according to RPM DB) |
These disk files need to be created and manipulated from a Linux system, preferably something that behaves like RHEL.
If your primary system is not running something RHEL-like, consider using a CentOS container:
docker run --privileged -v=$(pwd)/disks:/disks:rw -v=/dev:/dev -it centos:7 sh
Using this bare-bones Docker image, you may need to install additional packages such as e2fsprogs, ntfs-3g, and ntfsprogs.
The smallest a disk must be to format for ext4 is roughly 70 KiB. The following commands assume you want the smallest possible disk for that format.
Create the file and give it at least one partition:
dd bs=1024 count=70 if=/dev/zero of=/disks/my_new_disk
fdisk -C1 /disks/my_new_disk
Use a loop device to attach the disk and format it:
losetup -D # detach any existing loop devices
losetup -P /dev/loop0 /disks/my_new_disk
mkfs.ext4 /dev/loop0p1
Note: The default number of i-nodes on a 70K ext4 disk is 16. For additional inodes, use the -N option with the mkfs.ext4 command.
Mount the new partition and put things in it:
rm -rf /mnt/my_new_disk; mkdir -p /mnt/my_new_disk
mount -t auto,rw /dev/loop0p1 /mnt/my_new_disk
# write files into /mnt/my_new_disk
Unmount and detach the device when done:
umount /mnt/my_new_disk
losetup -D
When creating these minimal filesystems, you may need to create an RPM DB for houndigrade inspection purposes. Unfortunately, the RPM DB directory in a stock RHEL OS install is about 50 MB. If you find and download a small Red Hat-signed RPM (e.g. tree), however, you can install it locally against a nonstandard path to generate a minimal RPM DB.
For example:
rm -rf /tmp/rpmdb; mkdir /tmp/rpmdb
rpm --dbpath=/tmp/rpmdb --nodeps -i tree-1.7.0-15.el8.x86_64.rpm
rpm -qa --dbpath=/tmp/rpmdb/ --qf '%{NAME} %{SIGPGP:pgpsig}\n'
# tree RSA/SHA256, Wed Nov 7 17:20:51 2018, Key ID 199e2f91fd431d51
# "199e2f91fd431d51" is a key houndigrade will match for Red Hat
du -sh /tmp/rpmdb/
# 520K /tmp/rpmdb/