SLES 12 — google-compute-engine-init — multiple vulnerabilities (3 CVEs) — patch and remediation guide

📖 ~4 min read  •  Source: SUSE advisory SUSE-SU-2020:1934-1 (see also SUSE bugzilla)

Related CVEs: CVE-2020-8903 CVE-2020-8907 CVE-2020-8933

Upstream summary: A vulnerability in Google Cloud Platform's guest-oslogin versions between 20190304 and 20200507 allows a user that is only granted the role "roles/compute.osLogin" to escalate privileges to root. Using their membership to the "adm" group, users with this role are able to read the DHCP XID from the systemd journal. Using the DHCP XID, it is then possible to set the IP address and hostname of the instance to any value, which is then stored in /etc/hosts. An attacker can then po

Table of contents

1. Symptom & Impact
2. Environment & Reproduction
3. Root Cause Analysis
4. Quick Triage
5. Step-by-Step Diagnosis
6. Solution – Primary Fix
7. Solution – Alternative Approaches
8. Verification & Acceptance Criteria
9. Rollback Plan
10. Prevention & Hardening
11. Related Errors & Cross-Refs
12. References & Further Reading

Symptom & Impact

On SLES 12 hosts that have google-compute-engine-init installed, administrators report behaviour consistent with SUSE advisory SUSE-SU-2020:1934-1: zypper patch-check lists open patches, services backed by google-compute-engine-init fail or restart unexpectedly, AppArmor profile warnings appear in journalctl -k — and for security-rated advisories the host is exposed to the vulnerability set above. Impact ranges from a single service-restart loop to wider availability incidents whenever google-compute-engine-init sits on the serving path.

Environment & Reproduction

Reproduction targets SLES 12. Confirm release, registration, and installed package:

cat /etc/os-release
SUSEConnect --status-text
SUSEConnect --list-extensions 2>/dev/null | head -30
rpm -q google-compute-engine-init
zypper info google-compute-engine-init | head -20

Trigger the workflow that exposes google-compute-engine-init — multiple vulnerabilities (3 CVEs) — patch and remediation guide while collecting:

sudo journalctl -u google-compute-engine-init -b --no-pager | tail -200
sudo journalctl -xe --no-pager | tail -200
sudo tail -200 /var/log/zypp/history
sudo tail -200 /var/log/audit/audit.log
# For SUSE support, bundle evidence with supportconfig:
sudo supportconfig -R /var/tmp -B google-compute-engine-init

Root Cause Analysis

Root cause is documented in SUSE advisory SUSE-SU-2020:1934-1. SUSE security maintainers shipped fixes in the corresponding google-compute-engine-init update for SLES 12; running an outdated build leaves the host exposed to the failure modes described in the advisory. Correlate zypper history with system logs:

sudo zypper history | grep google-compute-engine-init
sudo zypper history --since='-7 days' | tail -40
sudo journalctl -k | grep -i apparmor | tail -100
cat /proc/sys/kernel/tainted   # non-zero = tainted kernel / out-of-tree modules

Quick Triage

Run these on SLES 12 to capture the current state of google-compute-engine-init:

rpm -q google-compute-engine-init                              # installed NVR
rpm -V google-compute-engine-init                              # verify shipped files
sudo zypper patch-check                    # open patches
sudo zypper lp -r SUSE-SLE-Server-12-* 2>/dev/null | head
systemctl --failed --no-pager
sudo firewall-cmd --list-all 2>/dev/null || sudo SuSEfirewall2 status 2>/dev/null
sudo aa-status                              # AppArmor profiles
# If google-compute-engine-init ships a systemd unit (unit name may differ from pkg name, e.g.
# bind→named, postgresql-server→postgresql, php-fpm→php-fpm):
systemctl list-unit-files | grep -i google | head

Step-by-Step Diagnosis

1. List failed systemd units.

   systemctl --failed --no-pager

2. Tail the journal for google-compute-engine-init and the system bus.

   sudo journalctl -u google-compute-engine-init -f --no-pager
   sudo journalctl -xe -f --no-pager

3. Inspect firewall posture. This release uses firewalld; SuSEfirewall2 may still be present on SLES 12 GA.

   sudo firewall-cmd --list-all-zones
   sudo SuSEfirewall2 status 2>/dev/null   # legacy, only present on early SLES 12
   sudo iptables -L -n -v | head -30

4. Surface AppArmor denials and switch the profile to complain mode if needed.

   sudo journalctl -k | grep -i 'apparmor="DENIED"' | tail -30
   sudo aa-status
   sudo aa-complain /etc/apparmor.d/usr.sbin.google-compute-engine-init 2>/dev/null || true

5. Verify google-compute-engine-init integrity and reinstall if anything is altered.

   sudo rpm -V google-compute-engine-init
   sudo zypper verify
   sudo zypper install --force google-compute-engine-init

6. Correlate findings with /var/log/zypp/history, zypper history, and SUSE advisory SUSE-SU-2020:1934-1 to pin the change that introduced google-compute-engine-init — multiple vulnerabilities (3 CVEs) — patch and remediation guide.

Solution – Primary Fix

Apply the corrective zypper transaction referenced by SUSE advisory SUSE-SU-2020:1934-1, then reload affected systemd units:

sudo zypper ref                        # refresh repos
sudo zypper -n patch                   # apply ALL open patches (recommended)
# Or target a single package:
sudo zypper -n update google-compute-engine-init
sudo systemctl daemon-reload
# Unit name may differ from pkg name; check first:
systemctl list-unit-files | grep -i google | head
sudo systemctl restart google-compute-engine-init
rpm -q google-compute-engine-init                           # confirm new NVR
systemctl is-active google-compute-engine-init 2>/dev/null  # confirm running (if a unit exists)

For kernel / glibc / systemd / openssl advisories a reboot is required (or SLE Live Patching where licensed):

sudo zypper ps -s                      # services using deleted libs
sudo systemctl reboot                  # or: sudo shutdown -r now
# SUSE Live Patching (kgraft / klp) avoids reboot for kernel CVEs:
sudo zypper install -y kernel-livepatch-$(uname -r | tr - _)
klp -v patches                         # active livepatches

Solution – Alternative Approaches

If the primary patch is not viable, choose from these:

• Roll back via Snapper (Btrfs snapshots taken automatically before zypper transactions on SLES 12):

  sudo snapper list
  sudo snapper undochange <pre>..<post>  # diff between two snapshot numbers
  sudo snapper rollback <pre>            # boot the host into the chosen snapshot

• Lock the package so zypper cannot upgrade it:

  sudo zypper al google-compute-engine-init                   # add lock
  zypper ll | grep google-compute-engine-init                 # list locks
  sudo zypper rl google-compute-engine-init                   # remove lock

• Install an older NVR if a regression is suspected:

  zypper se -s google-compute-engine-init                     # show all available versions
  sudo zypper install --oldpackage google-compute-engine-init-<older-NVR>

• If SuSEfirewall2 is still in use (rare on modern SLES 12), migrate to firewalld:

  sudo zypper install -y firewalld
  sudo systemctl disable --now SuSEfirewall2
  sudo systemctl enable --now firewalld

• Disable the AppArmor profile briefly to confirm policy is the cause, then re-enable:

  sudo aa-disable /etc/apparmor.d/usr.sbin.google-compute-engine-init
  # reproduce, capture denials in the journal:
  sudo journalctl -k | grep apparmor | tail
  sudo aa-enforce /etc/apparmor.d/usr.sbin.google-compute-engine-init

• Where SLE Live Patching is licensed, apply kernel fixes without reboot:

  klp -v patches                         # active livepatches
  sudo zypper install -y kernel-livepatch-$(uname -r | tr - _)

Verification & Acceptance Criteria

All of these should pass after the fix:

rpm -q google-compute-engine-init                                            # expected fixed NVR
sudo zypper patch-check                                  # 0 critical patches outstanding
systemctl is-active google-compute-engine-init 2>/dev/null
sudo journalctl -u google-compute-engine-init --since "5 minutes ago" --no-pager | grep -iE "error|fail" || echo OK
sudo firewall-cmd --list-services
sudo aa-status | head -5
sudo zypper ps -s                                        # any services still using deleted libs

The original reproduction for google-compute-engine-init — multiple vulnerabilities (3 CVEs) — patch and remediation guide must not trigger across two consecutive runs.

Rollback Plan

Capture state before any change:

rpm -qa > /root/rpm-pre.txt
sudo zypper history list > /root/zypper-history-pre.txt
# Snapper takes pre/post snapshots automatically on Btrfs root.
sudo snapper create -d 'pre-patch-google-compute-engine-init'   # explicit named snapshot
sudo snapper list | head

To revert if the patch is bad:

# Preferred on Btrfs root — boot the prior snapshot:
sudo snapper rollback <snapshot-id>
sudo systemctl reboot
# Or downgrade just the package:
sudo zypper install --oldpackage google-compute-engine-init-<older-NVR>
sudo systemctl daemon-reload
sudo systemctl restart google-compute-engine-init
# Custom security policy cleanup:
sudo apparmor_parser -R /etc/apparmor.d/usr.sbin.google-compute-engine-init

Prevention & Hardening

Reduce the chance of this recurring on SLES 12:

• Enable automatic patch installation:

  sudo zypper install -y zypper-automatic
  sudo systemctl enable --now zypper-automatic.timer
  # Or use YaST: yast2 online_update_configuration

• Subscribe to sle-security-updates and watch suse.com/support/update.

• Mirror through SUSE Manager or RMT (Repository Mirroring Tool) for controlled rollouts:

  sudo zypper install -y rmt-server rmt-cli
  sudo rmt-cli sync
  sudo rmt-cli products enable SLES/12/x86_64

• Lock sensitive packages so they cannot be auto-upgraded:

  sudo zypper al google-compute-engine-init

• Ensure Snapper is enabled on the root subvolume and pre/post hooks run for every zypper transaction:

  sudo snapper -c root get-config | head
  # Default zypper plugin: /usr/lib/zypp/plugins/commit/snapper.zypp-commit-plugin

• Monitor file integrity with AIDE:

  sudo zypper install -y aide
  sudo aide --init && sudo mv /var/lib/aide/aide.db.new /var/lib/aide/aide.db
  sudo aide --check

• Subscribe to SUSE Live Patching so kernel CVEs can be remediated without reboot:

  sudo SUSEConnect -p sle-module-live-patching/12.0/x86_64
  sudo zypper install -y kernel-livepatch-$(uname -r | tr - _)
  klp -v patches

• Keep AppArmor profiles in enforce; review /etc/apparmor.d/ after every package upgrade.

• Apply CIS SUSE Linux Enterprise Server Benchmark hardening.

Related Errors & Cross-Refs

Issues that commonly surface alongside google-compute-engine-init — multiple vulnerabilities (3 CVEs) — patch and remediation guide: zypper lock contention, systemd unit ordering cycles, AppArmor denials, firewalld zone drift, and kernel taint flags. Useful triage:

sudo zypper ps -s
systemd-analyze critical-chain
sudo journalctl -k | grep apparmor | tail
sudo firewall-cmd --get-active-zones
cat /proc/sys/kernel/tainted

View all sles-12 tutorials on the Tutorials Hub →

Browse all common problems & solutions on the Tutorials Hub.

References & Further Reading

Primary reference: SUSE advisory SUSE-SU-2020:1934-1 (see also SUSE bugzilla). Manual pages useful on SLES 12:

man zypper
man zypper.conf
man systemctl
man journalctl
man firewall-cmd
man snapper
man apparmor
man aa-status
man SUSEConnect
man klp

Other resources: SUSE Linux Enterprise Server 12 documentation, suse.com/security, SUSE security blog, and per-package notes in /usr/share/doc/packages/google-compute-engine-init/ for components implicated in google-compute-engine-init — multiple vulnerabilities (3 CVEs) — patch and remediation guide.