2
0
mirror of https://github.com/xcat2/confluent.git synced 2026-11-19 08:52:20 +00:00
Commit Graph

3563 Commits

Author SHA1 Message Date
Markus Hilger 903a9f2c9a Add deployment.installedargs for per-node installed kernel arguments
Kernel arguments for the installed OS could only be set per profile,
with installedargs in profile.yaml. deployment.installedargs sets them
per node or group, appended after the ones of the profile.

The deploycfg served to the node carries the value as installedargs,
and the installers add it wherever they add the profile installedargs:
the kickstart bootloader line on EL and RHVH, the AutoYaST and Agama
bootloader settings on SUSE, and GRUB_CMDLINE_LINUX on Debian and
Ubuntu. Diskless images installed to disk add it to GRUB_CMDLINE_LINUX,
and on EL9 and EL10 also to /etc/kernel/cmdline and the boot loader
entries.

deploycfg is no longer wrapped at 80 columns, since the scripts read
each value from its single line.
2026-11-18 23:44:04 +01:00
Jarrod Johnson 86f6c5f1f5 Merge pull request #391 from Obihoernchen/feat/node-kernelargs
Per-node kernel arguments at network boot
2026-11-18 16:17:47 -05:00
Jarrod Johnson 447d86739e Merge pull request #396 from Obihoernchen/feat/httpboot6-ipxe-shim
Use the iPXE shim for IPv6 UEFI HTTP boot
2026-11-18 16:09:47 -05:00
Jarrod Johnson 192eda00a5 Merge pull request #394 from Obihoernchen/fix/ipxe-shim-refresh
Fix ipxe shim refresh
2026-11-18 16:06:08 -05:00
Jarrod Johnson ba45554de1 Merge pull request #395 from Obihoernchen/fix/neighutil-addr-done
Stop the address dump at a done message anywhere in the datagram
2026-11-18 16:04:52 -05:00
Jarrod Johnson 631d7fa680 Relax watchdog timing
Some systems during startup especially may take a hit that crosses the threshold.

Also, some IO hangs in select environments during certain activity can pause a while.

Heartbeat 4 times per interval instead of 2.
2026-11-18 15:33:53 -05:00
Jarrod Johnson be48cb28ce Merge remote-tracking branch 'xcat/master' 2026-11-18 12:18:13 -05:00
Jarrod Johnson 00f62d4506 Fix some functions with chassis-only bmcs
Some infrastructure equipment models as a chassis with no system, handle this case better more generically.
2026-11-18 12:17:48 -05:00
Markus Hilger e6c5210699 Name the cause in the kernelargs HTTP boot warnings
UEFI HTTP boot goes through the iPXE shim once osdeploy initialize has
installed it, so the boot.img fallback warning points to that, for IPv4
and IPv6.
2026-11-18 13:30:05 +01:00
Markus Hilger 8d09bb5b0d Add deployment.kernelargs for per-node kernel arguments
Kernel arguments could only be set per profile, baked into the static
boot files of the profile. Nodes or groups that need extra arguments,
such as a different console or driver options, required a copy of the
whole profile.

When deployment.kernelargs is set, the iPXE boot offer now points at the
existing /confluent-api/boot/by-node endpoint, which serves the profile
boot.ipxe or boot.ipxe.shim with the node arguments appended to the
kernel line and the paths made absolute. UEFI HTTP boot takes the same
path, as it chainloads the iPXE shim when that is installed. A UEFI
HTTP boot that loads the static boot.img instead logs that the node
arguments are not applied.

Nodes without the attribute keep receiving the static profile files.
2026-11-18 13:30:01 +01:00
Markus Hilger a328123357 Ignore DHCPv6 requests sent to another server
A client sends its Request to the server whose Advertise it picked,
named by the server identifier. confluent answered every Request, so
with several DHCPv6 servers on the segment, such as the members of a
collective, all of them replied to it.
2026-11-18 13:06:14 +01:00
Markus Hilger a8f18b6955 Stop the address dump at a done message anywhere in the datagram
get_my_addresses only looked at the first netlink message of each
datagram for NLMSG_DONE. When the kernel sends the last addresses and
the done message in one datagram, as EL10 does for an IPv6 dump, the
loop waited for another datagram that never came. reply_dhcp6 asks for
the IPv6 addresses of the interface, so every DHCPv6 boot request went
unanswered.

A failed dump ends with an error message and no done message, so stop
at an error message too.
2026-11-18 13:06:00 +01:00
Markus Hilger be3623c551 Use the iPXE shim for IPv6 UEFI HTTP boot
IPv6 UEFI HTTP boot always served boot.img. Offer the iPXE shim when
the profile has boot.ipxe.shim and the shim of the booting
architecture is installed, as the IPv4 path does.

iPXE then sends its own DHCPv6 request with user class iPXE and a
PXEClient vendor class, which was refused as IPv6 PXE. Answer it with
the boot.ipxe.shim URL, or boot.ipxe when the profile has no shim
script. With only the other configuration flag in the router
advertisement, iPXE sends an Information-request, which was not
handled at all and carries no IA_NA. Handle it and only assign an
address when the request asks for one.

Before it starts the shim, the firmware releases the address and
retransmits the Release while it is unanswered, which delayed the
handoff by over a minute. Answer a Release sent to this server.
2026-11-18 11:22:22 +01:00
Markus Hilger aaaf4e3ea8 Install the HTTP boot iPXE shim on every osdeploy initialize run
The shim copy ran after the early exit of initialize, so -p, -k, -l
or -g on their own did not install it. It also checked for the aarch64
shim as ipxe-aarch64-shim.efi, a name it never writes, so it copied
that one on every run instead of only when it was missing.
2026-11-17 23:20:28 +01:00
Markus Hilger 2ba8db30ef Check the iPXE shim of the booting architecture for HTTP boot
UEFI HTTP boot chainloads the iPXE shim only when it is installed, but
the check always looked for the x86_64 shim. An aarch64 node on a
server with only the x86_64 shim was offered the missing aarch64 one
and failed to boot, and on a server with only the aarch64 shim it got
the static boot.img instead.
2026-11-17 21:05:33 +01:00
Jarrod Johnson f72ce3f98a Fix import of Ubuntu 26.04.1 arm media 2026-11-17 14:08:25 -05:00
Markus Hilger d7a4af65b1 Tolerate a null system UUID in Redfish inventory
A Giga Computing BMC reports the system UUID as null, which made
nodeinventory fail with 'NoneType' object has no attribute 'lower'.
Treat a null UUID like a missing one.
2026-11-17 16:41:50 +01:00
Markus Hilger aa250bb91b Handle IPMI sessions torn down while a response is pending
A SOL keepalive reply arriving after the console was closed failed with
an AttributeError on the released session. A privilege level request on
a session that was logged out or given up on during login raised
IpmiException out of the packet handler into the event loop shared by
all sessions. Ignore the late keepalive reply, and report the failed
login to its waiters instead.
2026-11-17 16:27:52 +01:00
Markus Hilger 75a5e75063 Log IPMI plugin stack traces to the trace log
traceback.print_exc() went to stderr, where the log redirect tags every
line with its caller, traceback.py, which made the traces unreadable.
2026-11-17 16:27:52 +01:00
Markus Hilger e73a42ca5e Report request errors from console sessions like other requests
A request sent from within a console session, such as the power state
check confetty makes, went through a handler that logged every failure
as an unexpected 500 with a stack trace. A node with an unreachable BMC
filled the trace log with TargetEndpointUnreachable. Share the error
reporting of the main request loop, so the client gets the real error
code and only unexpected failures are traced.
2026-11-17 16:27:52 +01:00
Markus Hilger cb729147e7 Tolerate credserver clients that disconnect early
A client that reset the connection or closed it before finishing the
exchange left a stack trace in the trace log (ConnectionResetError, or
an IndexError on an empty read). Read each field in full, and treat a
client going away as the end of that exchange rather than a server fault.
2026-11-17 16:27:52 +01:00
Jarrod Johnson 209d6ef2ed Fix undefined nodeinfo in the get_chassis_info flow. 2026-11-16 16:47:19 -05:00
Jarrod Johnson 246aae08fe Change secure http to ipxe as well
It is verified that iPXE leverages the UEFI CA and does not let untrusted connections.

So let it do https as well as http, relegating boot.img for virtual media instead.

This like the plain http boot works nicely with a broader set of UEFI.
2026-11-16 16:43:11 -05:00
Jarrod Johnson bd647ebfa4 Implement chassis information separate from system
Some implementations claim to have a system, but have useless data.

We can't tell programattically that the data is useless, so provide chassis for the user to decide.
2026-11-14 08:01:31 -05:00
Jarrod Johnson 334815e9a5 Catch the target url from a 403 extended message, if required 2026-11-13 10:00:38 -05:00
Jarrod Johnson d4c2072e3f Have nvos discovery handle LLA based onboarding 2026-11-13 09:36:31 -05:00
Jarrod Johnson dd700ad1a5 Add mac address to registered address, when possible 2026-11-13 09:04:07 -05:00
Jarrod Johnson 85ff47e684 Add singleton IPv6 support to nodediscover register
This allows otherwise queried ipv6 addresses to be checked one at a time.
2026-11-13 08:13:26 -05:00
Jarrod Johnson ad2de237bc Implement management of UEFI HTTPS certificates
On supported platforms, manage the UEFI HTTPS certificate store.  This can make it easier to manage HTTPS boot.
2026-11-12 10:17:48 -05:00
Jarrod Johnson 9d78d42071 Merge remote-tracking branch 'xcat/master' 2026-11-12 10:06:02 -05:00
Jarrod Johnson bec1ed1498 Merge pull request #383 from Obihoernchen/feat/confluent-attrib-helper
Add confluent_attrib helper for profile scripts
2026-11-12 07:40:14 -05:00
Markus Hilger c949359199 Return node groups from /self/myattribs
groups is stored as a plain list rather than a value dict, so the
value check skipped it and a node could not see its own groups. Return
it as a list, the same way ansible plays get it as host vars.
2026-11-11 23:02:02 +01:00
Jarrod Johnson 5523abc435 Merge pull request #382 from Obihoernchen/fix/nodedefine-shared-groups
Copy per-node attribute data in set_node_attributes
2026-11-11 13:02:08 -05:00
Markus Hilger 69a872c208 Copy per-node attribute data in set_node_attributes
nodedefine with a noderange passed one dict for all nodes, so they
shared one stored groups list. Later group membership changes then
updated inheritance for only the first node, leaving the others with
values inherited from groups they no longer belong to.
2026-11-11 18:51:56 +01:00
Jarrod Johnson 5990303661 Fix ARM/x86 UEFI HTTP boot
Make http location tree consistent with tftp
2026-11-10 14:09:37 -05:00
Jarrod Johnson 48f82777b4 Enable iPXE based http boot
When using non-https (when we would need UEFI certificate) and if we have secureboot friendly ipxe setup, use ipxe for http boot.

This avoids the tftp requirement and preserves secureboot, but uses material that UEFI is generally better equipped to handle.

For now, take care of the IPv4 path, IPv6 path contingent on testing.
2026-11-10 10:23:42 -05:00
Jarrod Johnson 52505e6f2d Improve SSDP responder concurrency
Bulk boot could get serialized too harshly. Particularly could hiccup on name lookup delay.

Take the penalties concurrently instead.
2026-11-06 17:28:05 -05:00
Jarrod Johnson 84faa9a1c7 Properly indicate bynode or not in message state 2026-11-06 10:08:19 -05:00
Jarrod Johnson 58d4c6a5a9 Allow API caller to request raw binary images.
This can make it more straightforward as an interface, and also defer the penalty for base64 processing.
2026-11-06 09:37:58 -05:00
Jarrod Johnson b0fff4a8fe Default to json, and add binary types for screenshot to come down in a more straightforward way 2026-11-06 08:44:46 -05:00
Markus Hilger 098d0a66c3 Enable ansible pipelining for plays by default
Without pipelining every task copies the module to the node over sftp,
then makes it executable, runs it and removes it, each a separate ssh
operation. With pipelining the module is fed to the interpreter over
the ssh session, so a task takes one operation and nothing is written
to the node's temporary directory.

Pipelining breaks become through sudo with requiretty, which rarely
applies here: plays connect as root and drop become unless become_user
is set. Setting ANSIBLE_PIPELINING or ANSIBLE_SSH_PIPELINING still
overrides this default.
2026-11-04 21:59:58 +01:00
Markus Hilger e668cdabf7 Keep ansible's default ssh_args when running plays
Setting ANSIBLE_SSH_ARGS replaces ansible's ssh_args instead of adding
to them, so plays ran without the default ControlMaster and
ControlPersist options. Every task then opened a new ssh connection
and authenticated again.

Start from ansible's default ssh_args when ANSIBLE_SSH_ARGS is not
set. An ANSIBLE_SSH_ARGS from the environment is still extended as
before.
2026-11-04 21:59:58 +01:00
Jarrod Johnson 5e45bc598b Merge pull request #367 from Obihoernchen/feat/ansible-node-attributes
Pass node attributes to profile ansible plays
2026-11-04 10:06:59 -05:00
Markus Hilger 9cbe3b91d8 Decode bytes attribute values outside of dicts too 2026-11-04 15:07:32 +01:00
Jarrod Johnson 830b6892d6 Merge pull request #369 from Obihoernchen/fix/runansible-hangs
Stop hanging on ansible play processes that fail early
2026-11-04 08:38:33 -05:00
Markus Hilger ec71d0b876 Stop hanging on ansible play processes that fail early
Drain the play process output while it runs, and stop waiting for its
feedback connection once it exits.
2026-11-04 00:23:33 +01:00
Markus Hilger 48b7dc62b5 Pass node attributes to profile ansible plays
Plays run by the deployment server get the target node's attributes in
the host variable 'confluent', nested at each dot of the name, e.g.
confluent['net']['ib0']['ipv4_address']. Secrets and unset attributes
are left out, and values are never templated.
2026-11-04 00:08:38 +01:00
Markus Hilger 1cad52284d Run profile ansible plays on ansible-core 2.19 and later
Newer ansible-core only takes a stdout callback name and no longer
templates untrusted YAML. Attach the results collector directly, load
plays trusted through the DataLoader, and report a crashed play
process as a failed task.
2026-11-03 23:58:43 +01:00
Jarrod Johnson 734ea9811a Also check if shim can be downloaded using tftp 2026-11-02 11:39:25 -05:00
Jarrod Johnson 89787671cf Merge pull request #338 from Obihoernchen/fix/ssh-agent-dir-owner
Hand /run/confluent/ssh to the confluent user
2026-10-31 09:21:18 -04:00