[gentoo-user] flask up/downgrade problem/question

[Jack]

Doing my usual emerge update this morning, portage insisted on 
downgrading several python packages, including flask.    It seems that 
flask-migrate was recently upgraded from 3.0.0 to 3.0.1, but it's DEPEND 
for flask was changed from >=flask-0.9 to local copy with the depend modified to >=flask-2, and it emerged with no 
error (also re-upgrading the other downgraded packages) but I have no 
idea whether I'm in for some later problems because of this.


I don't think a bug is in order, since the change seemed intentional, 
but I'd love to understand the reasoning.


Thanks for any pointers or background.

Jack

Re: [gentoo-user] cgroups and OpenRC: Am I doing it wrong?

[tastytea]

On 2021-07-15 09:58+0200 tastytea  wrote:

> www-apps/gitea often used all 4 of my CPU cores and slowed everything
> else down. So I attempted to limit CPU usage with cgroups:
> 
> I have set this in /etc/rc.conf:
> 
> rc_cgroup_mode="unified"
> rc_cgroup_controllers="cpu memory io pids"
> 
> And this in /etc/conf.d/gitea:
> 
> rc_cgroup_settings="
> cpu.shares 128
> cpu.cfs_period_us 100
> cpu.cfs_quota_us  100
> memory.limit_in_bytes 8G
> "

The error was that I used documentation for cgroups v1 instead of v2¹.
🤦The right values are:

rc_cgroup_settings="
cpu.max 100 100
memory.high 7G
memory.max 8G
"

¹ 

-- 
Get my PGP key with `gpg --locate-keys tasty...@tastytea.de` or at
.


pgpRHV5KnZVpQ.pgp
Description: Digitale Signatur von OpenPGP

Re: [gentoo-user] cgroups and OpenRC: Am I doing it wrong?

[tastytea]

On 2021-07-15 09:58+0200 tastytea  wrote:

> […] did I do someting wrong? How can I check if my settings are
> applied?

I've tried to set the nice value of my emacs daemon with 

rc_cgroup_settings="
cpu.weight.nice -10
"

/sys/fs/cgroup/emacs.tastytea/cpu.weight.nice shows -10 but htop says
that the nice value is 0.

-- 
Get my PGP key with `gpg --locate-keys tasty...@tastytea.de` or at
.


pgpt2i5zJfWCh.pgp
Description: Digitale Signatur von OpenPGP