xCAT/xcat-core
2
0
Fork 0
mirror of https://github.com/xcat2/xcat-core.git synced 2026-08-27 09:06:39 +00:00
Code Issues Releases Wiki Activity

Merge pull request #7730 from VersatusHPC/fix/enablekdump-staging

Browse Source 
fix(enablekdump): stage the NFS dump setup under a per-node directory
This commit is contained in:
Daniel Hilst
2026-08-24 14:36:48 -03:00
committed by GitHub
parent f2f96b67fc 5d3c40b5a2
commit 1e4242174f
3 changed files with 164 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/source/guides/admin-guides/manage_clusters/common/deployment/enable_kdump.rst
+5 -1
View File
@@ -132,7 +132,11 @@ Check your Operating System specific documentation for the path where the kernel
* **[RHELS6]** ::
<kdump_path>/var/crash/<node_ip>-<time>/
* **[RHELS7, RHELS8]** ::
<kdump_path>/<node hostname>/var/crash/<node_ip>-<time>/
* **[SLES11]** ::
<kdump_path>/<node hostname>/<date>
xCAT-test/unit/enablekdump_per_node.t
+137
View File
@@ -0,0 +1,137 @@
#!/usr/bin/env perl
use strict;
use warnings;
use File::Path qw(make_path);
use File::Temp qw(tempdir);
use FindBin;
use Test::More;
my $script = "$FindBin::Bin/../../xCAT/postscripts/enablekdump";
plan skip_all => 'enablekdump not found' unless -r $script;
# The postscript uses GNU sed -i, which behaves differently on BSD.
plan skip_all => 'postscript targets Linux nodes' unless $^O eq 'linux';
my $source = read_file($script);
# Run the RHEL NFS path of enablekdump against a scratch tree. The dump target
# is a local directory standing in for the mounted NFS export, so we can check
# what the postscript writes to it and to /etc/kdump.conf.
sub run_enablekdump {
my (%opt) = @_;
my $osver = $opt{osver};
my $node = $opt{node} || 'n01';
my $sysconfig = defined $opt{sysconfig} ? $opt{sysconfig}
: "KDUMP_COMMANDLINE=\"\"\nKDUMP_COMMANDLINE_APPEND=\"\"\n";
my $root = tempdir(CLEANUP => 1);
make_path("$root/etc/sysconfig", "$root/target", "$root/bin");
# /etc/sysconfig/kdump must exist for the in-place seds to land.
write_file("$root/etc/sysconfig/kdump", $sysconfig);
# xcatlib.sh is sourced; only restartservice is needed and is a no-op here.
write_file("$root/xcatlib.sh", "restartservice(){ :; }\n");
write_file("$root/bin/logger", "#!/bin/sh\nexit 0\n");
chmod 0755, "$root/bin/logger";
my $src = $source;
# Redirect everything the postscript touches into the scratch tree.
$src =~ s{/etc/kdump\.conf}{$root/etc/kdump.conf}g;
$src =~ s{/etc/sysconfig/kdump}{$root/etc/sysconfig/kdump}g;
$src =~ s{/etc/dracut\.conf}{$root/etc/dracut.conf}g;
$src =~ s{/tmp/dracut\.conf}{$root/tmp/dracut.conf}g;
# No real NFS server: make the version probe empty so the mount is skipped;
# the per-node mkdir and kdump.conf rendering still run against the target.
$src =~ s{/usr/sbin/rpcinfo}{/bin/false}g;
$src =~ s{/bin/mount}{true}g;
$src =~ s{/bin/umount}{true}g;
# Point the staging mount point at the scratch target that stands in for the
# mounted NFS export.
$src =~ s{/mnt/kdumpsetup}{$root/target}g;
write_file("$root/enablekdump", $src);
chmod 0755, "$root/enablekdump";
my $dump = "nfs://10.0.0.1/dumparea";
system(qq{cd '$root' && DUMP='$dump' XCAT='10.0.0.1:eth0' OSVER='$osver' }
. qq{ARCH='x86_64' NODE='$node' PATH="$root/bin:\$PATH" ./enablekdump >/dev/null 2>&1});
return {
root => $root,
target => "$root/target",
kdump_conf => read_file("$root/etc/kdump.conf"),
sysconfig => read_file("$root/etc/sysconfig/kdump"),
};
}
sub read_file {
my ($p) = @_;
return '' unless -e $p;
open my $fh, '<', $p or die "open $p: $!";
local $/;
return <$fh>;
}
sub write_file {
my ($p, $c) = @_;
open my $fh, '>', $p or die "open $p: $!";
print {$fh} $c;
close $fh;
return;
}
# --- RHEL 8: per-node, no shared-root writes -------------------------------
{
my $r = run_enablekdump(osver => 'rhels8.0', node => 'n01');
like($r->{kdump_conf}, qr{^path\s+/n01/var/crash$}m,
'kdump.conf points the dump path at the node subdirectory');
unlike($r->{kdump_conf}, qr{^path\s+/var/crash$}m,
'kdump.conf does not use the shared /var/crash path');
ok(-d "$r->{target}/n01/var/crash", 'the node subdirectory is created on the target');
ok(!-e "$r->{target}/var/crash", 'nothing is created at the shared export root');
ok(!-e "$r->{target}/proc", 'no dummy proc file is written on RHEL 8');
}
# --- RHEL 7: keeps the dracut workaround, but per-node ----------------------
{
my $r = run_enablekdump(osver => 'rhels7.9', node => 'n07');
like($r->{sysconfig}, qr{root=nfs:10\.0\.0\.1:/dumparea/n07},
'the RHEL 7 root= workaround points at the node subdirectory');
ok(-e "$r->{target}/n07/proc", 'the RHEL 7 dummy proc is written under the node subdirectory');
ok(!-e "$r->{target}/proc", 'the RHEL 7 dummy proc is not written at the shared root');
}
# --- RHEL 7 migration: a legacy shared root= is replaced, not kept ----------
# A node configured by the previous script carries root=nfs:<server>:<export>
# in KDUMP_COMMANDLINE_APPEND. dracut takes the last root= on the command
# line, so leaving the legacy value behind would defeat the migration.
{
my $legacy = qq{KDUMP_COMMANDLINE=""\n}
. qq{KDUMP_COMMANDLINE_APPEND="root=nfs:10.0.0.1:/dumparea rd.neednet=1 rootflags=nofail foo-root=bar rd.foo.root=bar"\n};
my $r = run_enablekdump(osver => 'rhels7.9', node => 'n07',
sysconfig => $legacy);
my ($append) = $r->{sysconfig} =~ m{^KDUMP_COMMANDLINE_APPEND="([^"]*)"}m;
my @roots = grep { /^root=/ } split ' ', defined $append ? $append : '';
is(scalar @roots, 1, 'exactly one root= remains after migrating a legacy config');
is($roots[0], 'root=nfs:10.0.0.1:/dumparea/n07',
'the remaining root= points at the node subdirectory');
like($append, qr{(?:^|\s)rd\.neednet=1(?:\s|$)},
'unrelated options on the legacy line are preserved');
like($append, qr{(?:^|\s)rootflags=nofail(?:\s|$)},
'rootflags= is not mistaken for a root= token');
like($append, qr{(?:^|\s)foo-root=bar(?:\s|$)},
'a root= suffix after a dash is not stripped');
like($append, qr{(?:^|\s)rd\.foo\.root=bar(?:\s|$)},
'a root= suffix after a dot is not stripped');
# Re-running against its own output must not stack another root=.
my $r2 = run_enablekdump(osver => 'rhels7.9', node => 'n07',
sysconfig => $r->{sysconfig});
my ($append2) = $r2->{sysconfig} =~ m{^KDUMP_COMMANDLINE_APPEND="([^"]*)"}m;
is($append2, $append, 'a second run leaves KDUMP_COMMANDLINE_APPEND unchanged');
}
done_testing();
xCAT/postscripts/enablekdump
+22 -17
View File
@@ -92,16 +92,13 @@ if [ ! -z "$DUMP" ]; then
KDIP=${XCAT%:*}
fi
# workaround for RHEL6
# xCAT sets NODE in the postscript environment; fall back to the short
# hostname so a per-node dump path is still used if run by hand.
[ -z "$NODE" ] && NODE=$(hostname -s)
# the $KDIP:$KDPATH directory will be used to generate the initrd for kdump service
MOUNTPATH=""
if (pmatch $OSVER "*6\.*"); then
MOUNTPATH="/tmp"
elif (pmatch $OSVER "*[78]\.*"); then
MOUNTPATH="/mnt"
else
MOUNTPATH="/var/tmp"
fi
MOUNTPATH="/mnt/kdumpsetup"
mkdir -p $MOUNTPATH
if [ "$KDPROTO" = "nfs" ]; then
if (pmatch $OSVER "sle*") || (pmatch $OSVER "suse*") || [ -f /etc/SuSE-release ] || [ -f /etc/SUSE-brand ]; then
@@ -238,16 +235,14 @@ EOF
else
/bin/echo "nfs server is not available"
fi
[ -d $MOUNTPATH/var/crash ] || mkdir -p $MOUNTPATH/var/crash
#The initramfs used in kdump does not need "root", however, the initramfs refused to continue
#if no valid "root" provided in redhat7.1 kdump; As a workaround,we provide a fake "root=nfs:$KDIP:$KDPATH"
#with a dummy "proc" inside, which will fake "root=nfs:$KDIP:$KDPATH" as a valid root directory
[ -e $MOUNTPATH/proc ] || echo "Dummy file: fake the /proc to pass the checking of 'root=' inside dracut-cmdline " > $MOUNTPATH/proc
# Create only this node's own subdirectory on the dump target.
# kdump checks the dump path exists when it builds the initrd, so
# it must exist now, at deploy time.
mkdir -p $MOUNTPATH/$NODE/var/crash
echo "nfs $KDIP:$KDPATH" > /etc/kdump.conf
echo "default shell" >> /etc/kdump.conf
echo "path /var/crash" >> /etc/kdump.conf
echo "path /$NODE/var/crash" >> /etc/kdump.conf
echo "core_collector makedumpfile -c --message-level 1 -d 31" >> /etc/kdump.conf
#strip "xcat" out of the initramfs for kdump
echo "dracut_args --omit \"xcat\"" >> /etc/kdump.conf
@@ -264,7 +259,17 @@ EOF
done
sed -i "s#^[\t ]*KDUMP_COMMANDLINE=\"#KDUMP_COMMANDLINE=\"$kdumpcmdline#" /etc/sysconfig/kdump
if (pmatch $OSVER "rhel7*") || (pmatch $OSVER "rhels7*"); then
sed -i "s#^[\t ]*KDUMP_COMMANDLINE_APPEND=\"#KDUMP_COMMANDLINE_APPEND=\"root=nfs:$KDIP:$KDPATH #" /etc/sysconfig/kdump
# The RHEL7 kdump initramfs refuses to continue unless root=
# names a valid filesystem, checked by dracut-cmdline at crash
# time, so give this node's subdir a dummy /proc for it.
[ -e $MOUNTPATH/$NODE/proc ] || echo "xCAT: dummy /proc so RHEL7 dracut-cmdline accepts root=" > $MOUNTPATH/$NODE/proc
# Drop any root= a previous run left behind before adding this
# node's own: dracut takes the last root= on the command line,
# so a stale shared-root value would win over the new one. A
# root= token starts after the opening quote or whitespace,
# so suffixes of other options (foo-root=, rd.foo.root=) stay.
sed -i -e '/^[\t ]*KDUMP_COMMANDLINE_APPEND="/{:a;s/\([" \t]\)root=[^ \t"]*[ \t]*/\1/;ta}' \
-e "s#^[\t ]*KDUMP_COMMANDLINE_APPEND=\"#KDUMP_COMMANDLINE_APPEND=\"root=nfs:$KDIP:$KDPATH/$NODE #" /etc/sysconfig/kdump
fi
[ -f /etc/dracut.conf ] && mv /etc/dracut.conf /tmp/dracut.conf
restartservice kdump