different kernel versions for today's patch

[oceanwave]

Probably a silly question, but why do two cloudlinux 9 servers get different kernel versions for today's patch, both showing they have the latest available:
5.14.0-611.49.2
5.14.0-611.54.1

[bogdan.sh]

I believe what you're seeing is CloudLinux's staged rollout. New kernels don't go out to every server at once; they're released in waves across 16 internal "slots" so any regressions get caught early. Each server is pinned to one slot, so two boxes can legitimately see different "latest" kernels at the same time.

The lagging one will pick up 5.14.0-611.54.1 on its own once the rollout reaches its slot (usually within a day or two).

You can confirm which slot each server is on with:

grep -i rollout /etc/yum.repos.d/*.repo

or if you want them aligned right now, you can install the kernel directly:

dnf install kernel-5.14.0-611.54.1

[bogdan.sh]

I guess it's about KernelCare.
The two servers aren't getting "the same patch" - they're getting the latest patch for the kernel each one happens to be booted on. KernelCare patches are scoped per booted RPM kernel, not global.

uname -r reflects the RPM kernel installed at boot time.

KernelCare maintains an independent patch stream per (distro, booted kernel) pair. When kcarectl --update runs, the agent fetches the patch series matching the kernel that is currently booted, applies the newest patch in that series, and reports "up to date." Two servers on two different booted kernels both get a "latest available" - but those latests come from different patch series, so the resulting effective kernel string differs.


Quick verification commands on the servers themselves:

uname -r

fot booted RPM kernel

kcarectl --uname -r

for effective post-patch version

kcarectl --patch-info | head

for patch series + level applied

Directly related, why uname is showing old kernel.

[oceanwave]

I'm sorry, I should have specified, these servers don't have kernelcare.
Just updated yesterday with dnf upgrade and then dnf upgrade kernel and then dnf upgrade and was surprised that the latest kernel on one was different than the other.

[bogdan.sh]

I believe what you're seeing is CloudLinux's staged rollout. New kernels don't go out to every server at once; they're released in waves across 16 internal "slots" so any regressions get caught early. Each server is pinned to one slot, so two boxes can legitimately see different "latest" kernels at the same time.

The lagging one will pick up 5.14.0-611.54.1 on its own once the rollout reaches its slot (usually within a day or two).

You can confirm which slot each server is on with:

grep -i rollout /etc/yum.repos.d/*.repo

or if you want them aligned right now, you can install the kernel directly:

dnf install kernel-5.14.0-611.54.1