makedhcp using the Kea backend fails for any node whose mac table entry
uses the *NOIP* sentinel to mark a secondary NIC that has no IP address
(e.g. "mac1|mac2!*NOIP*|mac3!*NOIP*"). kea_node_reservations() and
kea_node_reservations6() split each NIC's hostname and pass it straight
to getipaddr(), so the literal string "*NOIP*" is treated as a host to
resolve. It cannot resolve, and the reservation is reported as unresolved
-- which aborts the whole makedhcp run, leaving the node with no Kea
reservation at all and therefore no DHCP lease.
This breaks provisioning on distros that use the Kea backend (e.g. EL10,
which has no ISC dhcp-server): the node never gets an address and is
unreachable, while the ISC backend has always handled *NOIP* and is
unaffected.
Skip *NOIP* NICs in both the DHCPv4 and DHCPv6 reservation builders, the
same way the ISC path does -- there is no address to reserve for them.
Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
PR #7677 added xCAT-server/share/xcat/netboot/rh/compute.rhels9.ppc64le.pkglist
with a stray empty line at the end of the file. The reviewer asked for it to be
removed before merge, but it slipped through. A trailing blank line in a pkglist
is meaningless and inconsistent with the rest of the netboot package lists, so
remove it.
Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
Boot URLs handed to nodes always carried an explicit :80 even when the HTTP
server listens on the default port. grub2.pm already omits the port when it is
80; make the xNBA, genesis and DHCP boot URLs consistent with it.
$portsuffix is empty when site.httpport is unset or 80 and ":<port>" otherwise,
so a non-default port is unchanged. Covers the ISC and Kea paths in dhcp.pm,
xnba.pm, mknb.pm, and the Kea xNBA client classes in BootPolicy.pm.
dhcp_boot_policy.t and mknb_xcatd_address.t follow the new URLs and gain cases
for a non-default and an unset port. The floating-address guard in
mknb_xcatd_address.t also drops :80 so that it keeps matching.
Recovered from the unmerged lenovobuild branch (cd61fd9d, a2f77ea9, 1035841b,
b7db0e2, 792fe6d).
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
When invoked with no node name, replaycons printed a terse "Please specify a
node name." Print the full usage statement instead, consistent with -h.
Also document, in the man page, that confluent log rotation renames older logs
and how to replay one by appending the date suffix seen in
/var/log/confluent/consoles/.
Recovered from the unmerged lenovobuild branch (originals 1c1b23fac, 48a92b7d4).
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
Running rspconfig against an mgt=ipmi node with no option fell through to the
generic "Unsupported command: rspconfig " message (with an empty option),
which does not tell the user what is wrong. Detect the empty option in
preprocess_request and report that options are required.
Recovered from the unmerged lenovobuild branch (original 08ae94bd).
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
nodestat -f runs fping to reach the nodes; if the fping package is not
installed the pipe yielded no output and nodestat silently returned an empty
result. Check that an fping binary exists before running it and emit an
explicit "must install fping" error otherwise.
Recovered from the unmerged lenovobuild branch (original f0c88182), reworked
to test for the binary directly: the original inferred fping's presence from
whether the pipe produced output, which false-positived (reporting fping
missing) when fping was present but every node was unresolvable -- caught in
lab validation.
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
The FRU key sort used only `$a <=> $b`, a numeric comparison that returns 0
for the non-numeric FRU field names, leaving them in hash-iteration order.
That order is randomized per process, so `rinv <node> all` printed its fields
in a different order on each run. Add `$a cmp $b` as a tiebreaker so the
output is stable.
Recovered from the unmerged lenovobuild branch (original 9cd3fac6).
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
systemd tooling run inside the image chroot has no bus to talk to and prints
"Failed to connect to bus: No such file or directory". The image builds
correctly, so say so at the end of the run instead of leaving it to be guessed.
Recovered from the unmerged lenovobuild branch (e372fb5b).
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
noderes.netboot is parsed as grub2-<protocol> and validated with
/^http|tftp$/, which reads as (^http)|(tftp$) and so also accepts values
such as https, httpx and xtftp.
A value that passes but is not exactly "http" then falls through to
set root=$grub2protocol,$serverip, which drops site.httpport, so the mistake
surfaces as a confusing grub2 failure at boot instead of the "Invalid netboot
method" error this check exists to give. grub2-https cannot work in any case,
since only "insmod http" is emitted.
Supported values are unchanged: grub2, grub2-http and grub2-tftp.
nodeset aborted with "Unable to find pxelinux.0 from syslinux" when neither
the copy under the tftp directory nor the one under XCATROOT existed. Only
legacy BIOS clients chain pxelinux and not every distribution still ships
syslinux, so a UEFI-only environment could not run nodeset at all.
Copy pxelinux.0 when it is available and warn, rather than fail, when it is not.
Recovered from the unmerged lenovobuild branch (4f0a5106), which skipped
silently; warn so the missing loader is still reported.
Co-authored-by: Jarrod Johnson <10814490+jjohnson42@users.noreply.github.com>
Vinícius Ferrão
Daniel Hilst