Veeam warns of Software Appliance boot failures on AMD EPYC hosts

Veeam has issued a warning to its customers regarding boot failures impacting Software Appliance virtual machines running on host systems powered by AMD EPYC processors. These appliances are pre-configured Linux-based VMs designed to run the Veeam Backup & Replication suite.

According to the advisory, the issue affects Veeam Software Appliance version 13.1 and later. The root cause lies within the Linux kernel used by the appliance, which runs integrated health and cryptographic checks in FIPS mode during early boot.

FIPS mode checks and kernel panic on AMD EPYC

To comply with enterprise and government security frameworks, the underlying Linux kernel executes a jitterentropy health test in FIPS mode during early boot. However, this test intermittently fails on certain AMD EPYC processors.

Affected processors include specific models such as AMD EPYC 7262, 7302, and 7F72. When the test fails during a reboot or kernel update, the kernel halts the boot process entirely and crashes with a kernel panic error.

Mitigation steps and recommended workarounds

To mitigate this issue, Veeam advises IT administrators to redeploy the affected virtual appliance to a Linux host equipped with a different, unaffected processor model. If hardware migration is impossible, admins should disable automatic updates and apply updates manually.

If a kernel panic occurs, simply restarting the guest operating system will not recover the VM; administrators must power off the virtual machine and power it back on. Once successfully booted, normal backup operations resume without further issues.

Additional MPTCP vulnerability and enterprise impact

In a separate advisory, Veeam noted that some of its products are impacted by a critical Linux kernel vulnerability (CVE-2026-98432) in the IPv4 TCP networking stack. This flaw could allow remote unauthenticated attackers to execute arbitrary commands as root on servers running MPTCP listeners.

Given that Veeam solutions are deployed by over 550,000 customers globally, including a vast majority of Fortune 500 corporations, addressing these boot stability issues and kernel patches promptly is vital for maintaining robust enterprise data protection infrastructures.