Security review: fix traversal, identity, injection and takeover paths
Findings from a full review (auth/API, CA and rack channel, config generation, integrations, UI/agents/scripts), each verified in code: - rack binary endpoint read any region file via arch=../.. (CA key) - render endpoint traversal; machine-reported MACs used as file paths - delegated intermediates could mint admin/rack identities and IPs; identities now require root-issued certs, intermediates are ServerAuth-only with both name types constrained - rack certs were ServerAuth for rack-chosen SANs: now client-only; racks accept only the region's own cert; join checks the CA fingerprint; unbound join codes cannot take an existing name; job results bound to the rack that took them - rack self-update: region signs binaries with the CA key, the root unit verifies against the join-time CA in a root-owned dir - dnsmasq reads a checked, root-owned copy limited to generated options - commissioning could overwrite deployed machines; MAC takeover of PXE - injection into dnsmasq/cloud-init/PXE/udev/DNS via names and machine-reported strings; stored XSS in the UI; CSP and cross-origin protection; deploy agent shell injection and seed traversal - Redfish requests confined to the BMC's /redfish/ tree - admin_http defaults to loopback; body limits and timeouts; secrets off command lines Also: IPAM claim/reconcile race, sync stopping on one error, watchdog race, release wiping the wrong disk, SAN LUNs as auto targets. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
1 parent
6e4be736ec
commit
a549f01ef5
28 files changed
+782
-154
No files matched your search
+41
-3
@@ -161,6 +161,12 @@ func (s *Store) FinishJob(ctx context.Context, id, state, errMsg string) error {
|
||||
return err
|
||||
}
|
||||
|
||||
// FailPendingJob fails a job that has not started; false when it started meanwhile.
|
||||
func (s *Store) FailPendingJob(ctx context.Context, id, errMsg string) (bool, error) {
|
||||
tag, err := s.pool.Exec(ctx, `UPDATE jobs SET state='failed', finished=now(), started=coalesce(started, now()), error=nullif($2,'') WHERE id=$1 AND state='pending'`, id, errMsg)
|
||||
return err == nil && tag.RowsAffected() == 1, err
|
||||
}
|
||||
|
||||
func (s *Store) SetDeployed(ctx context.Context, id, os, version string) error {
|
||||
_, err := s.pool.Exec(ctx, `UPDATE machines SET state='deployed', os=$2, os_version=$3, deployed_at=now(), booted_at=NULL, last_seen=now() WHERE id=$1`, id, os, version)
|
||||
return err
|
||||
@@ -176,6 +182,9 @@ func (s *Store) SetReleased(ctx context.Context, id string) error {
|
||||
}
|
||||
|
||||
func (s *Store) SetNodeAddr(ctx context.Context, id, name, mgmt, ib string) error {
|
||||
if name != "" && !validDomain(name) {
|
||||
return fmt.Errorf("node name %q is not a host name (letters, digits, hyphens)", name)
|
||||
}
|
||||
_, err := s.pool.Exec(ctx, `UPDATE machines SET node_name=$2, mgmt_ip=$3, ib_ip=$4 WHERE id=$1`, id, name, nilIfEmpty(mgmt), nilIfEmpty(ib))
|
||||
return err
|
||||
}
|
||||
@@ -213,6 +222,9 @@ func inventoryDisks(m *Machine) []invDisk {
|
||||
|
||||
// pickDisk: the smallest non-removable, non-USB disk that fits the image (boot/OS disks are
|
||||
// usually the small ones on servers; data disks stay untouched). Explicit choice wins.
|
||||
// remoteTran: transports of disks that are not the machine's own (lsblk TRAN).
|
||||
var remoteTran = map[string]bool{"usb": true, "iscsi": true, "fc": true, "fcoe": true, "srp": true, "iser": true}
|
||||
|
||||
func pickDisk(disks []invDisk, want string, need int64) (*invDisk, error) {
|
||||
if want != "" {
|
||||
for i := range disks {
|
||||
@@ -230,7 +242,7 @@ func pickDisk(disks []invDisk, want string, need int64) (*invDisk, error) {
|
||||
}
|
||||
var cands []invDisk
|
||||
for _, d := range disks {
|
||||
if d.Removable || d.Tran == "usb" || d.Size < need {
|
||||
if d.Removable || remoteTran[d.Tran] || d.Size < need { // never a USB stick or a SAN LUN by itself
|
||||
continue
|
||||
}
|
||||
cands = append(cands, d)
|
||||
@@ -468,6 +480,20 @@ func (a *App) handleDeploy(w http.ResponseWriter, r *http.Request) {
|
||||
fail(w, 400, fmt.Errorf("os is required"))
|
||||
return
|
||||
}
|
||||
req.Hostname = strings.ToLower(req.Hostname)
|
||||
if req.Hostname != "" && !validDomain(req.Hostname) {
|
||||
fail(w, 400, fmt.Errorf("hostname %q is not a host name (letters, digits, hyphens)", req.Hostname))
|
||||
return
|
||||
}
|
||||
// refuse before anything is allocated or renamed
|
||||
if m.Profile != nil {
|
||||
fail(w, 409, fmt.Errorf("machine is a stateless node (profile %s); unassign it first", *m.Profile))
|
||||
return
|
||||
}
|
||||
if j, err := a.store.ActiveJob(r.Context(), m.ID); err == nil {
|
||||
fail(w, 409, fmt.Errorf("job %s (%s, %s) is still active; cancel it first", j.ID, j.Kind, j.State))
|
||||
return
|
||||
}
|
||||
meta, err := osResolve(a.cfg.OSDir, req.OS, req.Version)
|
||||
if err != nil {
|
||||
fail(w, 404, err)
|
||||
@@ -550,6 +576,17 @@ func (a *App) handleRelease(w http.ResponseWriter, r *http.Request) {
|
||||
if req.Wipe == "" {
|
||||
req.Wipe = "quick"
|
||||
}
|
||||
if req.Disk == "" { // the disk the OS went to, not "the smallest one"
|
||||
if jobs, err := a.store.Jobs(r.Context(), m.ID, 50); err == nil {
|
||||
for _, j := range jobs {
|
||||
var sp DeploySpec
|
||||
if j.Kind == "deploy" && j.State == "done" && json.Unmarshal(j.Spec, &sp) == nil && sp.Disk != "" {
|
||||
req.Disk = sp.Disk
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
disk, err := pickDisk(inventoryDisks(m), req.Disk, 0)
|
||||
if err != nil {
|
||||
fail(w, 409, err)
|
||||
@@ -775,8 +812,9 @@ func (a *App) jobWatchdog() {
|
||||
continue
|
||||
}
|
||||
msg := fmt.Sprintf("the machine did not start the job within %s: check its power, that it PXE-boots on the provisioning network, and its console for a hung POST (events show the power command's result; job_start_timeout raises the limit)", limit)
|
||||
_ = a.store.FinishJob(ctx, j.ID, "failed", msg)
|
||||
a.afterJob(ctx, m, j, "failed", msg)
|
||||
if ok, _ := a.store.FailPendingJob(ctx, j.ID, msg); ok { // not if it reported progress since the snapshot
|
||||
a.afterJob(ctx, m, j, "failed", msg)
|
||||
}
|
||||
}
|
||||
}
|
||||
cancel()
|
||||
|
||||
Reference in new issue
Block a user