Not able to enable the memory cgroup in cachyos

General
1

When running Elasticsearch docker container, it fails with the following cgroup error.

java.lang.NullPointerException: Cannot invoke "jdk.internal.platform.CgroupInfo.getMountPoint()" because "anyController" is null

A quick search shows that it might be related to memory cgroup not enabled. /proc/cgroups does not show memory subsystem.

cat /proc/cgroups | column -t
#subsys_name  hierarchy  num_cgroups  enabled
cpu           0          190          1
cpuacct       0          190          1
blkio         0          190          1
devices       0          190          1
freezer       0          190          1
net_cls       0          190          1
perf_event    0          190          1
net_prio      0          190          1
hugetlb       0          190          1
pids          0          190          1
rdma          0          190          1
misc          0          190          1

So, I added some additional kernel params to /etc/sdboot-manage.conf:

cgroup_enable=memory cgroup_memory=1

then ran sudo sdboot-manage gen and rebooted. I verified that I see the new kernel params in /prod/cmdline.

But, the memory subsystem is still not listed in /proc/cgroups and elasticsearch docker container still fails to start.

Any pointers to get the memory cgroup enabled would be much appreciated. Thanks.

2

Possibly your kernel was built without support. There is also the important distinction of whether its attempting to use cgroup v1 or v2.

What is the output of

zgrep 'CONFIG_MEMCG' /proc/config.gz

and

uname -a
3

Hi @NiceTM

zgrep 'CONFIG_MEMCG' /proc/config.gz

CONFIG_MEMCG=y
# CONFIG_MEMCG_V1 is not set

uname -a

Linux workstation 6.12.0-2-cachyos #1 SMP PREEMPT_DYNAMIC Mon, 18 Nov 2024 18:02:00 +0000 x86_64 GNU/Linux
4

Hi @NiceTM any ideas here?

5

Actually, never mind. Switching to a newer version of the elasticsearch docker image fixed the error for me.