xCAT/xcat-core
2
0
Fork 0
mirror of https://github.com/xcat2/xcat-core.git synced 2026-09-05 04:27:55 +00:00
Code Issues Releases Wiki Activity

fix(xcat-core): bound the boot-flip exchange, and fail when the install server will not resolve

Browse Source 
Two ways the Subiquity install could fail without saying anything useful.

The boot flip read from the install monitor with no timeout. A monitor that accepts the
connection and then never answers -- which #7759 shows is a state it gets into -- blocked the read
forever, and with it the Subiquity late-command and the install. The retry loop could not help:
it never reached the retry. Both reads now take -t 10, so five attempts are bounded at roughly two
minutes and end in the failure that is already logged. The regression test stands up a listener
that accepts and holds the connection; removing the timeouts fails it.

mkinstall resolved the install server for nfsroot and fell back to the name when that failed:
"getipaddr($instserver) || $instserver". The name is the original defect -- klibc's nfsmount
cannot resolve one -- so the node panicked "can't parse IP address" at boot, on the node, with
nothing reported on the management node. The management node knows at template time, so it says
so there and skips the node, as the other unrecoverable per-node conditions in this routine do.

Signed-off-by: Daniel Hilst <392820+dhilst@users.noreply.github.com>
This commit is contained in:
Daniel Hilst
2026-08-28 10:51:55 -03:00
parent 0f12faa0fc
commit bca12ca29d
3 changed files with 28 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
xCAT-server/lib/xcat/plugins/debian.pm
+9 -1
View File
@@ -1032,7 +1032,15 @@ sub mkinstall {
my $kcmdline = "nofb utf8 auto xcatd=" . $instserver;
if (using_subiquity($os,$tmplfile)) {
my $nfsip = xCAT::NetworkUtils->getipaddr($instserver) || $instserver;
# Fail here rather than handing casper a name: klibc's nfsmount cannot resolve
# one, so the node would panic "can't parse IP address" at boot, on the node,
# with nothing said on the management node.
my $nfsip = xCAT::NetworkUtils->getipaddr($instserver);
unless ($nfsip) {
xCAT::MsgUtils->report_node_error($callback, $node,
"Could not resolve the install server '$instserver' to an address. The Ubuntu live installer mounts its root with klibc nfsmount, which cannot resolve names, so nfsroot must be an address.");
next;
}
$kcmdline = subiquity_kcmdline($kcmdline, $nfsip, $pkgdir, $instserver, $httpport, $node);
} else {
$kcmdline .= " url=http://${instserver}:$httpport/install/autoinst/$node";
xCAT-server/share/xcat/install/ubuntu/compute.subiquity.tmpl
+4 -2
View File
@@ -110,7 +110,9 @@ autoinstall:
# Flip the node to local-disk boot, or it PXE-loops back into the installer on reboot.
# "next" is the request xcatd's install monitor answers with "nodeset <node> next". Send it
# from the live installer, retried, and log a failure rather than silently reinstalling. A
# monitor that died and never came back is a separate problem; see #7759.
- ['bash', '-c', 'xm=#XCATVAR:XCATMASTER#; ok=0; for i in 1 2 3 4 5; do if exec 3<>/dev/tcp/$xm/3002; then read -r r <&3 || true; printf "next\n" >&3; if read -r r <&3; then ok=1; fi; exec 3>&- 3<&-; [ "$ok" = 1 ] && break; fi; sleep 5; done; if [ "$ok" != 1 ]; then echo "xcat: FAILED to flip $(hostname) to local-disk boot via $xm:3002; the node will PXE back into the installer" >>/target/var/log/xcat/xcat.log; fi; exit 0']
# monitor that died and never came back is a separate problem; see #7759 -- but one that
# accepts the connection and then never answers would block a bare read forever and hang the
# install here, so both reads are bounded.
- ['bash', '-c', 'xm=#XCATVAR:XCATMASTER#; ok=0; for i in 1 2 3 4 5; do if exec 3<>/dev/tcp/$xm/3002; then read -r -t 10 r <&3 || true; printf "next\n" >&3; if read -r -t 10 r <&3; then ok=1; fi; exec 3>&- 3<&-; [ "$ok" = 1 ] && break; fi; sleep 5; done; if [ "$ok" != 1 ]; then echo "xcat: FAILED to flip $(hostname) to local-disk boot via $xm:3002; the node will PXE back into the installer" >>/target/var/log/xcat/xcat.log; fi; exit 0']
error-commands:
- tar -c --ignore-failed-read --transform='s/^/#HOSTNAME#-logs\//' /var/crash /var/log/installer /tmp/pre-install.log /autoinstall.yaml 2>/dev/null |nc -l 8080
xCAT-test/unit/ubuntu_subiquity_boot_flip.t
+15 -1
View File
@@ -61,6 +61,7 @@ sub run_flip {
for (1 .. $opt{listen}) {
my $c = $srv->accept() or last;
$c->autoflush(1);
if ($opt{mute}) { sleep 600; close $c; next } # accept and hold, never answer
print {$c} "ready\n";
my $line = <$c>;
print {$seen} $line if defined $line;
@@ -75,7 +76,8 @@ sub run_flip {
# The installer would hang here if the exchange ever blocked, so bound it.
# the no-listener case prints "Connection refused" by design
my $rc = system("timeout 25 bash -c \Q$script\E 2>/dev/null");
my $cap = $opt{cap} || 25;
my $rc = system("timeout $cap bash -c \Q$script\E 2>/dev/null");
my $timed_out = (($rc >> 8) == 124);
if ($pid) { kill 'TERM', $pid; waitpid($pid, 0) }
@@ -114,6 +116,18 @@ sub run_flip {
'a connection without an acknowledgement counts as a failure, not a success');
}
# --- a monitor that accepts and never answers must not hang the install ----
# The late-command runs inside Subiquity: a bare read on a socket that is open but silent blocks
# forever and the install never finishes. #7759 fixes the monitor dying; this bounds the wait.
{
# Five attempts, each bounded by two 10s reads plus the retry pause: ~2 minutes worst case.
my $r = run_flip(listen => 1, mute => 1, cap => 200);
ok(!$r->{timed_out},
'a monitor that accepts but never replies does not hang the late-command');
like($r->{log}, qr/FAILED to flip/,
'it is recorded as a failed flip rather than waiting indefinitely');
}
# --- the command retries rather than giving up on the first refusal --------
{
# Answer only on a later connection: the flip must still succeed.