FS#33620 - [procps-ng] Settings related to net bridge in sysctl.conf are not applied at boot

Attached to Project: Arch Linux
Opened by Dan Ziemba (zman0900) - Monday, 28 January 2013, 23:30 GMT
Last edited by Gaetan Bisson (vesath) - Sunday, 09 June 2013, 11:58 GMT
Task Type Bug Report
Category Packages: Core
Status Closed
Assigned To Gaetan Bisson (vesath)
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:
I am using arch as a xen dom0 and have a bridge set up for my domUs. I added the following to /etc/sysctl.conf to avoid iptables dealing with bridge traffic:

# Disable netfilter on bridges.
net.bridge.bridge-nf-call-ip6tables = 0
net.bridge.bridge-nf-call-iptables = 0
net.bridge.bridge-nf-call-arptables = 0

On reboot, all other settings in sysctl.conf take effect, but nothing related to the bridge does. Running `systemctl restart systemd-sysctl` applies the settings. This seems to be becuase the bridge module does not get loaded until later. I'm not sure if this would be considered a bug, but this does work in other Linuxes so I'm reporting it here.

Additional info:
* package version(s)
kernel 3.7.4-1-ARCH
systemd 197

* config and/or log files etc.
# cat /etc/network.d/br0
INTERFACE="br0"
CONNECTION="bridge"
DESCRIPTION="Bridge connection"
BRIDGE_INTERFACES="enp5s0"
IP="dhcp"
## sets forward delay time
#FWD_DELAY=0
## sets max age of hello message
#MAX_AGE=10
POST_UP="brctl stp br0 on"


Steps to reproduce:
Configure network bridge br0
Enable at boot with `systemctl enable netcfg@br0`
Add setting to /etc/sysctl.conf to disable netfilter on bridges
Reboot
sysctl -a
This task depends upon

Closed by  Gaetan Bisson (vesath)
Sunday, 09 June 2013, 11:58 GMT
Reason for closing:  Upstream
Additional comments about closing:  udev (see Dave's comment)
Comment by Danny Navarro (vimes656) - Tuesday, 05 March 2013, 10:00 GMT
The same is happening to me with 'net.ipv4.tcp_allowed_congestion_control = cubic reno lp'
Comment by Gaetan Bisson (vesath) - Saturday, 04 May 2013, 23:45 GMT
Why is procps-ng at fault here? This seems to me like a kernel bug: the sysctl settings are passed on to the kernel, and the latter ignores them when there is no bridge. At any rate, this should be reported upstream: there is nothing we can do about it packaging-wise.
Comment by Dave Reisner (falconindy) - Sunday, 05 May 2013, 00:33 GMT
udev rules are supposed to re-apply relevant sysctl rules as network devices appear, but it's racy and doesn't always work.

There really isn't much of a solution to this.

P.S. I urge you to use /etc/sysctl.d/*.conf rather than /etc/sysctl.conf.

Loading...