Veeam has warned that its Software Appliance virtual machines (VMs) can fail to boot on hosts powered by certain AMD EPYC processors. The appliances may crash with a kernel panic during the early boot process after a Linux kernel fails a FIPS-mode jitter entropy health check. The issue affects Veeam Software Appliance versions 13.1 and later.
Veeam Software Appliance is a preconfigured, Linux-based virtual machine that enables administrators to run the Veeam Backup & Replication software suite on Linux devices without manually installing dependencies or configuring the software.
To meet corporate and government security requirements, the underlying Linux kernel performs integrated sanity and cryptographic checks during startup when operating in Federal Information Processing Standards (FIPS) mode. These checks include tests for jitter entropy, which is used to help provide secure randomness.
In an advisory published Wednesday, Veeam said the virtual machines intermittently fail to start after reboots, including reboots following updates that install new kernels, on hosts with certain AMD EPYC processors.
“Veeam Software Appliance (Linux-based Veeam Backup & Replication) virtual machines running on hosts with AMD EPYC processors intermittently fail to start after reboots, including reboots after updates that install new kernels,” the company said.
According to Veeam, the Linux kernel used by the appliances runs a jitterentropy health test in FIPS mode during initial boot. The test intermittently fails on certain AMD EPYC processors, including the AMD EPYC 7262, 7302 and 7F72. When that happens, the kernel halts the startup process with a kernel panic.
Users on affected systems may see the following error:
Kernel panic - Not in sync: jitterentropy: Initialization failed on host that is not compliant with requirements: 9.
Veeam recommends moving the appliance or controlling updates
Veeam advises IT administrators to redeploy the appliance to a Linux host with a different processor to mitigate the problem.
If moving the appliance to a system with an unaffected processor is not possible, administrators should disable automatic updates on the appliance and install updates manually while monitoring VM restarts.
“If a kernel panic occurs, power off the virtual machine and power it on again. Rebooting the guest OS will not recover the virtual machine,” Veeam said. “After the appliance boots successfully, no further errors occur and Veeam Backup & Replication operations are not affected.”
Veeam also addresses a critical Linux kernel vulnerability
Earlier today, Veeam reported that some of its products are affected by CVE-2026-98432, a critical vulnerability in the Linux kernel’s IPv4 TCP networking code.
The flaw allows a remote, unauthenticated attacker to execute commands as root on a server running a Multipath TCP (MPTCP) listener. Veeam has fixed the vulnerability in some affected components and is distributing patched kernels to affected systems through a private fix or Veeam Updater.
Veeam says its products are used by more than 550,000 customers, including 74% of the Global 2,000 and 82% of the Fortune 500.
Source: www.bleepingcomputer.com