Bug 250060

Summary:	shutting down of bonding interface loops forever under special conditions
Product:	Red Hat Enterprise Linux 4	Reporter:	David Kostal <david.kostal>
Component:	initscripts	Assignee:	initscripts Maintenance Team <initscripts-maint-list>
Status:	CLOSED DUPLICATE	QA Contact:	Brock Organ <borgan>
Severity:	low	Docs Contact:
Priority:	low
Version:	4.5	CC:	notting
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2009-03-18 21:19:37 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:

Description David Kostal 2007-07-30 07:40:45 UTC

Description of problem:
Shuttind down (ifdown) of bonding interface with one link down can result in
infinite loop. If complete reboot/poweroff is being ewxecuted, the machine will
not shut down.

Version-Release number of selected component (if applicable):
initscripts-7.93.29.EL-1

How reproducible:
Allways

Steps to Reproduce:
1. Configure bond0 and bond1 (balance-alb  in my case), with eth0,eth1 in bond0
and eth2,eth3 in bond1. eth2 has no physical link, other network cards are
connected.
2. Start up bond0 and bond1
3. Try to shutdown bond1 (or do complete poweroff)
  
Actual results:
Machine appears to hang while shutting down bond1. In fact it's executing
"ifdown eth2", "ifdown eth3", "ifdown eth2",... in a loop.

Expected results:
bond1 should shut down instantly.

Additional info:
It's is possible that this error appears with bond0 configured only if there was
also one (probably the first) slave without link. However it was tested only
with bond0 and bond1 configuration by me and therefore also the steps to
reproduce include both bond0 and bond1.

The scenario also caused "freeze" of the system during shutdown and machine had
to be resetted with a button.

If both slave devices are with link, there is no problem and everything works as
expected.

The problem was fixed instantly after removing HWADDR= lines from slave devices'
configurations. Unfortunately this also caused renaming of the eth devices on
next boot/after reloading ethernet card driver.

Comment 1 David Kostal 2007-08-22 06:16:23 UTC

Were you able to reproduce it? Are there any news on this issue? Thanks.

Comment 2 Bill Nottingham 2007-08-29 17:09:52 UTC

No updates as of now.

Comment 3 Bill Nottingham 2009-03-18 21:19:37 UTC

I believe this will be fixed in a future RHEL 4 update.

*** This bug has been marked as a duplicate of bug 448837 ***