We want to make you aware of an upstream Linux kernel issue that can prevent EPI2ME workflows from starting on some Ubuntu 24.04 (“Noble”) systems. This is of particular note, as this is the operating system on which Oxford Nanopore Technologies’ sequencing systems are based.
First and foremost:
The underlying problem originates from changes in the Linux kernel and Java, which are outside our direct control.
This is an extended form article of that originally published in the Nanopore Community
Under certain Ubuntu 24.04 configurations, EPI2ME workflows fail to start with an error similar to:
ERROR ~ Cannot invoke "jdk.internal.platform.CgroupInfo.getMountPoint()" because "anyController" is null
This happens when:
We have verified that the issue is resolved by upgrading to the fixed 6.14.0-36-generic kernel from the package
linux-generic-hwe-24.04 version 6.14.0-36.36~24.04.1
EPI2ME workflows run as normal on systems using this kernel.
You are likely affected if all of the following are true:
CgroupInfo and anyController being null.
You are unlikely to be affected if:
You can check your Ubuntu version by running the following command in a console:
lsb_release -d
and looking for something like:
Description: Ubuntu 24.04.3 LTS
This will confirm the version of the Ubuntu operating system. The OS kernel version can be checked using:
uname -r
If you see:
The characteristic error of the incompatible OS kernels is seeing the following error in workflow logs:
N E X T F L O W ~ version 23.04.2 ... ERROR ~ Cannot invoke "jdk.internal.platform.CgroupInfo.getMountPoint()" because "anyController" is null
We strongly recommend moving to the fixed HWE kernel if your environment allows it. On Ubuntu 24.04, as a user with administrative rights:
sudo apt update sudo apt install --install-recommends linux-generic-hwe-24.04sudo reboot
After rebooting verify that:
uname -r
shows the updated kernel:
6.14.0-36-generic
In our internal testing, 6.14.0-36-generic (from linux-generic-hwe-24.04 6.14.0-36.36~24.04.1) fully resolves the issue.
If EPI2ME is currently running without issues on your Ubuntu 24.04 system:
If you are on 6.8.x or another kernel version known to work, make a note of it. Avoid jumping to intermediate 6.12–6.14 kernels that do not include the 6.14.0-36 fix.uname -r
This issue is fundamentally an interoperability bug between newer Linux kernels (≥ 6.12), the cgroups v2 control-group interface, and the Java runtime used by Nextflow (and therefore by EPI2ME workflows).
In version 6.12, the Linux kernel changed how certain cgroup controllers (notably cpuset and memory) are exposed under /proc/cgroups and /sys/fs/cgroup.
Many Java runtimes use these files to detect available CPU and memory constraints inside containers.
On affected kernels the cpuset and memory controllers no longer appear where the Java runtime expects them. This causes Java’s internal cgroup-detection code to encounter a missing controller and throw a fatal error. All this happens before Nextflow starts an analysis workflow, so EPI2ME workflows fail immediately at startup.
We have confirmed that:
We expect to move EPI2ME to a newer Nextflow version in a future release so that we can take advantage of the OpenJDK 21 fix. The underlying OpenJDK issue has been fixed in the Java 21 development branch and is targeted for a future Java 21.0.10 update. That release is expected as part of the normal quarterly Java update cycle in January 2026.
In the meantime, the most practical and reliable fix is the kernel update to 6.14.0-36-generic.
We recognise that kernel and Java issues are not what you want to be dealing with when you are trying to run experiments. We are now validating new kernel and Java combinations against MinKNOW and EPI2ME, and updating our internal QA processes and platform guidance to reduce the chance of similar surprises in the future.
If you believe you are affected and need help applying any of the corrective actions above, please contact our support team with:
lsb_release -d or cat /etc/os-release)uname -r)We’ll work with you to get your system back to a stable, working state as quickly as possible.
Related Links
