This reverts commit be6c7a3794.
They do not exist in Leap/SLE 16 installer initrd.
The diskless hook keeps its copies. imgutil's installkernel instmods
all four, so they are in that image and the calls do work there.
Building a SUSE 16 image from SLE media failed every package with
"key ID fec28eaf09d9ea69: NOKEY". Leap publishes that key as
gpg-pubkey-*.asc, which the existing glob picks up; SLE publishes the
same key only as repodata/repomd.xml.key, so nothing was imported.
15 media carries both spellings, so this changes nothing there.
udev's kmod builtin dlopens libkmod, and dracut installs it from an
inst_libdir_file line in a module-setup.sh rather than by following
NEEDED. Which module carries that line moved: the dracut on SLE 16
media declares it only in 00systemd, which the diskless module set
never loads, so the image came up with no libkmod, udev autoloaded
nothing, and the guest reached the network scan with only loopback.
Leap's newer dracut also declares it in 95udev-rules, which base
depends on, which is why Leap was unaffected.
OpenSSH 10 splits each connection into sshd-session and that into
sshd-auth, so the initramfs sshd on 2222 could not serve a single
session. el10 added sshd-session for the same reason.
SLES 16 ships no ISC dhclient and nothing provides dhcp-client, so the
image could not be built from SLES media at all. Leap carries dhcpcd
too, so one client covers both. el10 made the same move when RHEL
dropped dhclient.
The deploycfg carries the literal 'null', which went into the profile as
a password hash, so the installed root account reported a usable
password instead of a locked one. 15 substitutes '!' for this; the sed
delimiter has to move off '!' to carry it.
The default Etc/UTC is not in the tzdata list agama validates against,
so the load failed and the install fell through to agama defaults and
still reported completion. Normalize it and halt if the load fails.
The attached-media branch could never be taken: the label pattern built
from os-release is opensuse-leap-16.0 or sles-16.0, while the media is
labelled Install-Leap-16.0-x86_64 and Install-SUSE-SLE-16-x86_64. Had it
matched, it would have written an inst.repo to the dracut cmdline that
agama does not read, and skipped inst.script entirely.
They stayed commented out when the hook was forked from el8, so an
IPoIB-only node had no path to the deploy server. The diskless hook
loads them already.
prechroot.sh had setupssh.sh commented out and copied only the keys
inline, so nodes came up without shosts.equiv, without the CA in
ssh_known_hosts and without a setuid ssh-keysign.
- the aarch64 osdeploy spec builds the stateful suse16 addons but its
diskless loop was never extended, so the aarch64 rpm shipped suse16
without suse16-diskless and a packed image got a dangling addons.cpio.
- imgutil's builddeb keeps its own copy of the directory list that
confluent_imgutil.spec.tmpl has, and it had learned about neither suse16
nor el10.
- gather_bootloader gained a /usr/share/efi fallback for shim on both
architectures but only for x86_64 on grub, so an aarch64 root found a
shim and then died copying grub.
Finally, rewriting repos.d file by file rather than copying the tree meant
a subdirectory or a file that is not valid UTF-8 aborted the build before
any package was installed, which also regressed SUSE 15. Pass anything
that is not a plain text repo definition through untouched and restore the
modes on the ones that are rewritten.
SuseHandler refused anything but 15.x. What 16 needed beyond widening it:
- its repo urls are written in terms of ${releasever}, which zypper
resolves from the target root's os-release, a file that does not exist
yet when the first packages go in
- its repos name a zypper service backed by a package-provided directory
the target root does not have, so zypper discarded every one of them
as an orphan
- there is no mkinitrd to work out which kernel to build for, and bare
dracut would build for the build host's running kernel
- the efi payloads moved out of /usr/lib64/efi, arping out of /usr/sbin,
nsswitch.conf and protocols under /usr/etc, and the presets enable
sshd already
- the module list predated virtio, so an image built for a KVM guest had
no network at all, and dm-crypt could not allocate a transform for the
encrypted image without the aes-xts modules
- urlmount still links libpthread, an empty stub since glibc 2.34 that
nothing else in the initramfs pulls in
Ported from suse15-diskless, with the differences 16 forces:
- dracut symlinks /lib/dracut/hooks to /var/lib/dracut/hooks, so a hook
shipped at the old path replaces the symlink with a directory
- there is no netconfig or /etc/sysconfig/network to hand the running
address to, so the initramfs writes a NetworkManager keyfile instead.
Without it NetworkManager claims the interface on its own terms and the
tethered root filesystem goes away with the old address, and confignet
never gets the chance to refine anything
- the discovery loop retries without a delay, so a link that takes a
moment to come up can exhaust all 30 tries before the first packet can
go anywhere. Keep asking, as the el9 hook already does
Without it osdeploy import cannot generate a profile at all:
generate_stock_profiles opens profile.yaml unguarded, and initprofile.sh
seds the label into it. The label substitution also still looked for
'sle 15'.
init_sdr assigned self._sdr before initialize() ran, so a failure left the
half built object in the cache. The next call saw a non-None _sdr and handed
back that partial repository rather than trying again.
The visible symptom is a first call raising and the second appearing to
succeed. The real cost is on a bmc where the read fails once: the client
keeps the incomplete sdr for the life of the session and every later sensor
lookup answers from it without complaint.
get_webclient falls off its end when /api/login answers anything but 200, so
it returned None. wc() passes that back, and thirty of the thirty-four call
sites use it unchecked, so a refused login arrived as "'NoneType' object has
no attribute 'grab_json_response'" from wherever it landed.
Raised where the failure is known, and with the status: 404 is firmware with
no web api, or a Redfish-only capture of one, while 401 is credentials it
will not take. Neither was distinguishable before.
The other four call sites are the inventory reads, and they always did check.
They answer partially when the web interface is out of reach, which is why
nodeinventory still says something useful. They ask through wc_if_available
now, so that tolerance is stated rather than resting on a None.
Two places where the code that exists to explain a failure fails instead, and
the caller is shown the second failure rather than the first.
_do_web_request builds its message from the response body when that body is
not the JSON error document the spec asks for. An html 404 page is exactly
that, the body is bytes, and str + bytes raised TypeError. The status and the
body never reached anyone.
LenovoFirmwareConfig raised a bare Exception when python-lxml and
python-eficompressor are absent. Confluent has no handler for one, so a
missing dependency showed as "Unexpected Error" and hid a message that
already said what to install.
Neither changes what fails, only what the caller is told.
Same function, one line above. EthernetInterfaces is optional, and when it is
absent the None went into a request and raised TypeError from the url library
rather than saying what was missing.
Kept beside the ambiguous case because the two are one question asked twice.
Dedicated plus shared is how BMCs are built, so several NICs is ordinary.
_get_bmc_nic_url only reaches the count when the address the session came in
on matches none of them, which is what a tunnel or a NAT does, and it then
raised the bare PyghmiException base class.
Confluent has no handler for the base class, so it fell through to the
generic one and showed "Unexpected Error" plus a traceback, for a machine
doing nothing wrong. UnsupportedFunctionality now, which the redfish plugin
already reports plainly and which stays inside PyghmiException.
The message said "does not have exactly one interface" without saying how
many, which ones, or what to do. Every caller takes a name, so it lists the
candidates, and the empty case reads differently from the ambiguous one.
A Redfish service may publish no Systems collection, and power and cooling
equipment does exactly that. sysurl is then None, and get_power and
get_bootdev handed it straight to a request, raising TypeError from inside
the url library.
sysinfo already guarded the same field. Both now ask through _system_url and
get a refusal naming what is missing. DMTF publish three services of this
shape, which is why they sit commented out in inventory-dmtf.yaml.
PCIeDevices and PCIeFunctions are optional, and get_health indexed both
without checking. _get_adp_urls in the same file already spells it
.get('PCIeDevices', []), so these two were the outliers.
Power and cooling equipment has no PCIe at all. The KeyError escaped the
health read and reached the user as "Unexpected Error" with a traceback
behind it. Reproduces offline against DMTF's public-rackmount1 mockup.
nodesensors and nodeconfig printed an error and exited 0, so
`nodesensors n1 && next-step` ran next-step after the read it guarded had
already failed.
nodesensors had three of these: the per-node branch never set the exit code,
the top-level branch beside it read `exitcode |= exitcode`, and a normal
return from main() fell off the end of the file. nodeconfig accumulates with
|= all through its read path except the last line, which assigned, so a
failed bmc configuration read was discarded by the system read after it.
The fixed branches now match how nodehealth and client.py spell the same
thing.
Markus Hilger
Jarrod Johnson
Jarrod Johnson