852b574d01
I *think* this is why our tests started failing recently. It seems somehow very unlikely to me though that we'd somehow managed to avoid `f` in the boot checksums until now, but without doing some math...it seems plausible. Closes: #1243 Approved by: jlebon
26 lines
858 B
Bash
Executable File
26 lines
858 B
Bash
Executable File
#!/bin/bash
|
|
set -xeuo pipefail
|
|
|
|
dn=$(cd $(dirname $0) && pwd)
|
|
. ${dn}/libcomposetest.sh
|
|
|
|
prepare_compose_test "bootlocation-new"
|
|
pysetjsonmember "boot_location" '"new"'
|
|
runcompose
|
|
echo "ok compose"
|
|
|
|
# Nothing in /boot (but it should exist)
|
|
ostree --repo=${repobuild} ls -R ${treeref} /boot > bootls.txt
|
|
cat >bootls-expected.txt <<EOF
|
|
d00755 0 0 0 /boot
|
|
EOF
|
|
diff -u bootls{-expected,}.txt
|
|
# Verify /usr/lib/ostree-boot
|
|
ostree --repo=${repobuild} ls -R ${treeref} /usr/lib/ostree-boot > bootls.txt
|
|
assert_file_has_content bootls.txt vmlinuz-
|
|
assert_file_has_content bootls.txt initramfs-
|
|
kver=$(grep /vmlinuz bootls.txt | sed -e 's,.*/vmlinuz-\(.*\)-[0-9a-f].*$,\1,')
|
|
# And use the kver to find the kernel in /usr/lib/modules
|
|
ostree --repo=${repobuild} ls ${treeref} /usr/lib/modules/${kver}/{vmlinuz,initramfs.img} >/dev/null
|
|
echo "ok boot location new"
|