Bug 497451

Summary:	vif0.0 not set correctly when ifcfg-eth0 has MTU value
Product:	Red Hat Enterprise Linux 5	Reporter:	matt
Component:	xen	Assignee:	Miroslav Rezanina <mrezanin>
Status:	CLOSED WORKSFORME	QA Contact:	Virtualization Bugs <virt-bugs>
Severity:	medium	Docs Contact:
Priority:	low
Version:	5.3	CC:	clalance, llim, matt, mrezanin, mshao, qwan, xen-maint, yuzhang
Target Milestone:	---
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2010-07-14 04:53:19 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:	514498

Description matt 2009-04-23 23:03:49 UTC

Description of problem:
When MTU value is defined in ifcfg-eth0 in Dom0 and xen kernel + associated startup scripts are used, the same MTU is propergated to peth0, xenbr0 and eth0 but not vif0.0. This leads to terrible network performance.
Setting MTU to the same value as the previously mentioned devices solves this issue.

Version-Release number of selected component (if applicable):
5.3

How reproducible:
Every time.

Steps to Reproduce:
1. Configure MTU=9000 in /etc/sysconfig/network-scripts/ifcfg-eth0 in Dom0
2. Reboot
3. Attempt large data transfer over, say, SMB
  
Actual results:
Dire network performance. Getting 2MB/s rather that 40MB/s.
As an aside, agitates issues with network under Vista, cause it to become unresponsive and require Disable/enable of network interface

Expected results:
Similar throughput to none Xen environment


Additional info:
Can be resolved by setting MTU manual on vif0.0
Testing has not yet been done for DomUs.

Comment 3 Miroslav Rezanina 2010-07-01 09:16:14 UTC

Retesting shows that MTU is not propagated to any other device than eth0 and performance is not degrading. 

What scripts do you use?? Can you give us xend-config.sxp for this case or retest it with rhel 5.5?

Comment 4 matt 2010-07-01 14:47:09 UTC

I do not have the environment to test this further.

Comment 8 Miroslav Rezanina 2010-07-14 04:53:19 UTC

As we are not able to reproduce the problem and reproduce does not have the enviroment to test, we close this as WORKSFORME. If you ever encounter this problem, feel free to reopen the bz providing the setup of enviroment you reproduced the problem - ifcfg script file, ifconfig output.