Bug 528822

Summary:	Honor DEFROUTE=yes\|no for all connection types
Product:	[Fedora] Fedora	Reporter:	Dan Williams <dcbw>
Component:	initscripts	Assignee:	Bill Nottingham <notting>
Status:	CLOSED RAWHIDE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	medium	Docs Contact:
Priority:	low
Version:	rawhide	CC:	fitzsim, gczarcinski, notting, rvokal
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:
Clones:	530209 (view as bug list)		Environment:
Last Closed:	2009-11-08 20:16:25 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	530209

Description Dan Williams 2009-10-13 21:13:31 UTC

There are various problems with GATEWAYDEV, specifically that it uses device names, and there's no way to sanely modify it from a GUI if the device you'd like to set as the gateway dev isn't plugged in, you don't have persistent device names, etc.  We shouldn't be requiring the user to type in a freeform text field what device name to use here.  Plus, if the NM connection you are editing isn't locked to a device, we dont' know the device name to use, since it's a per-connection setting and not global like GATEWAYDEV is.

So, the proposal is to expand usage of DEFROUTE in the ifcfg files.  If DEFROUTE=no, then 'ifup' for that ifcfg should not touch the default route.  That's done for PPP apparently, but not for the other device types.

NM will write out DEFROUTE to the ifcfg file based on the "Only use this connection for resources on its network" checkbox, and leave GATEWAYDEV alone.  NM will still honor GATEWAYDEV when its set though.

Comment 1 Dan Williams 2009-10-13 21:13:51 UTC

See https://bugzilla.redhat.com/show_bug.cgi?id=528281 for the NM side of this.

Comment 2 Bill Nottingham 2009-10-14 20:51:09 UTC

The assumption is that no DEFROUTE implies DEFROUTE=yes, correct?

Comment 3 Dan Williams 2009-10-15 15:31:03 UTC

(In reply to comment #2)
> The assumption is that no DEFROUTE implies DEFROUTE=yes, correct?  

Yes; that should be how it currently works in initscripts, and how NetworkManager works as well.

Obviously if DEFROUTE=yes for ifcfg-eth2 but GATEWAYDEV=eth0, then eth2 still doesn't get the default route.

Comment 4 Bill Nottingham 2009-10-21 21:19:27 UTC

http://git.fedorahosted.org/git/?p=initscripts.git;a=commitdiff;h=a23c3828e092315e71ba4d5e6befc060a4f2c07a

Will be in 9.02-1.

Comment 5 Gene Czarcinski 2009-11-08 20:16:25 UTC

code is there and looks correct ... 9.02-1 is in F12