Red Hat Bugzilla – Bug 70969
Network interface shutdowns randomly
Last modified: 2007-04-18 12:45:18 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.1b) Gecko/20020801
Description of problem:
Sometimes (mostly in night about 23h but it could be just accident) network
interface gets shutdown with this message in /var/log/messages:
Aug 5 23:29:42 localhost network: Shutting down interface eth0: succeeded
Aug 5 23:29:43 localhost network: Shutting down loopback interface:
Aug 5 23:29:43 localhost /etc/hotplug/net.agent: NET unregister event not
Aug 5 23:29:43 localhost kernel: PCI: Found IRQ 10 for device 00:0d.0
Aug 5 23:29:43 localhost kernel: 3c59x: Donald Becker and others.
Aug 5 23:29:43 localhost kernel: 00:0d.0: 3Com PCI 3c905C Tornado at
0xdc00. Vers LK1.1.16
This happens on two different machines (one using 3Com, other one using e1000 +
sis900). When I recompiled kernel (RH 7.3 2.4.18 + XFS patch) with different
options it stopped. I suspect that helped when I turned off "hotplug" options at
Version-Release number of selected component (if applicable):
Steps to Reproduce:
what exact kernel version are you using?
I'm using default RH 7.3 kernel (2.4.18-4) with XFS support
(2.4.18-4SGI_XFS_1.1). I'm using default kernel as it comes. When I recompile it
with some features disabled (like hotplug) it doesn't appear any more. As it is
production system I cannot do any more tests and on my test system this error
Please report bugs in the SGI Linux kernel to SGI, not us.
Sorry, but I disagree - this error seems to be connected to hotplug and/or
network part of the system. Even on one system I was forced stop using XFS (some
little bug in endian conversion causing file corruption when copying large
amounts of files and data) and this error was still occuring. Also when I left
XFS compiled in and turned off another parameters (like I mention - hotplug
support) the error disappeared.
I know it's almost impossible to reproduce this bug so I don't expect it to be
solved - just wanted to make a bugzilla record in case somebody else comes to
this problem also.