A set of tools to work with cgroup tree and processes classification/QoS, according to it.
More (of a bit outdated) info can be found here.
Main script there - cgconf - allows to use YAML like this to configure initial cgroup hierarcy:
path: /sys/fs/cgroup
defaults:
_tasks: root:wheel:664
_admin: root:wheel:644
freezer:
groups:
# base:
# _default: true # to put all pids here initially
# cpu.shares: 1000
# blkio.weight: 1000
user.slice:
cpu:
cfs_quota_us: 1_500_000
cfs_period_us: 1_000_000
tagged:
cave:
_tasks: root:paludisbuild
_admin: root:paludisbuild
cpu:
shares: 100
cfs_quota_us: 100_000
cfs_period_us: 250_000
blkio.weight: 100
memory.soft_limit_in_bytes: 2G
desktop:
roam:
_tasks: root:users
cpu.shares: 300
blkio.weight: 300
# memory.soft_limit_in_bytes: 2G
java:
_tasks: root:users
cpu.shares: 100
blkio.weight: 100
memory.soft_limit_in_bytes: 1G
(from laptop with dying fan, hence cpu bandwidth limits)
And then something like cgrc <tagged-name> <cmd> <args...> to run anything put
inside these.
Other tools allow waiting for one cgroup threads to finish before proceeding and put stuff running there on hold easily.
cgconf and cgrc turn out to be surprisingly still useful despite systemd adding knobs to control cgroup resource limits (but not all of them, and spread over lot of small files, which are pain if you need a big picture of e.g. weights) and systemd-run, which hides i/o of whatever it runs in systemd slices.
- check out peo3/cgroup-utils and whether I can rebase this stuff on top of it