Bug#940021: systemd: socket activiation leads to OOM situation due to slices not getting cleaned up

Michael Biebl biebl at debian.org
Sat Feb 1 11:25:19 GMT 2020


On Wed, 11 Sep 2019 11:29:11 +0200 =?utf-8?q?Julian_H=C3=BCbenthal?=
<julian at huebenthal.io> wrote:
> Package: systemd
> Version: 241-7~deb10u1
> Severity: normal
> 
> Dear Maintainer,
> 
> i currently troubleshoot an issue with my Check_MK instance shooting my LXC Container (on Proxmox 6) with Debian 10.1 Buster out of memory after around 10 hours. 
> I already tried to mitigate/work a round the situation with MaxConnectionsPerSource and TimeoutSec of systemd but the issue persists. 
> 
> Normally, as far as i understand the man pages regarding systemd.socket and systemd.service the service instance created by the socket should be get cleaned up. 
> What i observe with "systemd-cgtop" is there are staying many slices which are growing and using around 600.0K+ memory each and doesn't clean up. 
> Which leads to an OOM situation after some hours.  As i'm unable to find anything regarding clean up or a setting which should fix this i think it's a bug? 
> 
> I can't reproduce it on an Ubuntu 18.04, what i see there is that the slices getting cleaned up correctly. 
> I observe the same on a fresh Debian 10.1 VM on VMWare Workstation. 
> 
> I see many of the following entrys in "systemd-cgtop": 

Julian, Chris, assuming you can still reproduce this with v244 from
buster-backports, it would be great if you can file this upstream at
https://github.com/systemd/systemd/issues and report back with the issue
number.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <http://alioth-lists.debian.net/pipermail/pkg-systemd-maintainers/attachments/20200201/a0f68cf6/attachment.sig>


More information about the Pkg-systemd-maintainers mailing list