The riscv64 kernel was built without ACPI. Firmware that publishes ACPI
tables and no device tree - server-class RISC-V boards, and edk2 with
acpi=on - therefore left the kernel with no hardware description: it died
before console initialisation, so a node handed the image by grub2 went
silent at ExitBootServices and never reached the discovery action.
Enable ACPI, its error interfaces and the EFI stub the image is already
booted through. Rebuilt with these options, the same image on the same
firmware parses the RSDP, XSDT, FACP, DSDT, APIC, RHCT and SPCR tables,
reaches multi-user and starts discovering, and the kernel now also sees
the SMBIOS entry point the firmware exports.
The genesis boot script sleeps 20 seconds unconditionally with the comment
"wait for NIC initialization", but that wait exists solely for System P LE
(ppc64) hardware discovery -- the surrounding section is explicitly ppc64
only. Every x86_64 and aarch64 node therefore pays a needless 20s delay on
each genesis boot.
Move the sleep inside the `[[ ${ARCH} =~ ppc64 ]]` block so only ppc64 waits.
Applied to the base script and both dracut_105 (el, ubuntu) variants.
Recovered from the unmerged lenovobuild branch (original d0916d9d).
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
Building the genesis netboot image on a usr-merged host (openSUSE Leap
15, and increasingly EL) leaves the extracted root with /bin, /sbin,
/lib and /lib64 as real directories holding the same files that already
live under /usr/*. rpm then sees each binary twice (e.g. fs/bin/bash and
fs/usr/bin/bash) and aborts the install of xCAT-genesis-base with dozens
of file conflicts, so the package cannot be installed. After unpacking
the dracut image, fold those top-level directories into /usr and replace
them with relative symlinks, matching a normal usr-merged layout. The
loop is a no-op when the image already ships them as symlinks, so EL
builds are unaffected.
Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
A 2.17 -> 2.18 management-node upgrade installs xCAT-genesis-base 2.18 over the
2.14.5 payload that 2.17 shipped, and aborted in three ways that only surface on
the upgrade path (a fresh EL10 install never hit them):
1. usr/lib/dracut/hooks changed type: 2.14.5 shipped it as a real directory,
modern dracut (EL8+) makes it a symlink to ../../../var/lib/dracut/hooks. RPM
refuses to replace a directory with a symlink across an upgrade ("file ...
conflicts"). The %pretrans removal does not help because rpm computes the
conflict from package metadata, not the live filesystem. Materialize the
symlink back into a real directory (with the hook contents) in %install so the
payload type matches the installed 2.14.5 layout.
2. The %pretrans Lua used table.getn()/unpack(), removed in Lua 5.2+, so on
EL8/EL9 (and any upgrade where the dirs exist) the scriptlet died with
"attempt to call a nil value (field getn)". Use #t / table.unpack.
3. The %post ran "mknb" unconditionally; during a full "dnf update xCAT" xcatd is
stopped, so mknb cannot connect and exited non-zero, failing the whole
transaction. Tolerate the failure and leave the genesis-base-updated marker so
the image is rebuilt once xcatd is back.
The genesis-base deb version was frozen at an ancient value, below the lower
bound required by xcat-genesis-scripts (xcat-genesis-base-amd64 >= 2.13.10).
apt therefore refused to install it and its dependents. Bump the version to
track current xCAT so the dependency resolves.
Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
This fixes a problem where the dracut image was pulling artifacts from
the build host and not the mock image, e.g. building for EL9 in EL10.
Signed-off-by: Daniel Hilst Selli <392820+dhilst@users.noreply.github.com>
Vinícius Ferrão
Daniel Hilst
Abhishek S A
Markus Hilger
Markus Hilger
Mark Gurevich
Kilian Cavalotti