Bug 825160

Summary: Change network with netconsole loaded cause kernel panic
Product: Red Hat Enterprise Linux 6 Reporter: Guohua Ouyang <gouyang>
Component: ovirt-nodeAssignee: Mike Burns <mburns>
Status: CLOSED ERRATA QA Contact: Virtualization Bugs <virt-bugs>
Severity: high Docs Contact:
Priority: high    
Version: 6.3CC: acathrow, bsarathy, chchen, cshao, fdeutsch, gouyang, hadong, hambrose, huiwa, jboggs, leiwang, mburns, moli, ovirt-maint, sgordon, ycui, yuachen
Target Milestone: rcKeywords: TestOnly, ZStream
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: ovirt-node-2.5.0-3.el6 Doc Type: Bug Fix
Doc Text:
Previously, changing the network configuration of the Red Hat Enterprise Virtualization Hypervisor while netconsole was enabled resulted in a kernel panic ("Kernel panic - not syncing: Watchdog detected hard LOCKUP"). This kernel bug has now been fixed. The hypervisor has been updated to include a later kernel release that does not have this bug. Changing the network configuration of the Hypervisor with netconsole enabled will not result in a kernel panic.
 Story Points: ---
Clone Of:
: 839266 (view as bug list) Environment:
Last Closed: 2013-02-28 16:34:30 UTC Type: Bug
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: 839266    
Bug Blocks: 846431    
Attachments:
Description Flags
kernel_panic screenshot
none
A trace
none
netconsole kernel panic none

Description Guohua Ouyang 2012-05-25 09:25:59 UTC 
Description of problem:
After the network is up and netconsole is loaded, try change the network, such change network from dhcp to static or configure another nic, kernel panic will happen.
on screen can see "Kernel panic - not syncing: Watchdog detected hard
LOCKUP on cpu 6".

Version-Release number of selected component (if applicable):
rhev-hypervisor6-6.3-20120523.1.el6

How reproducible:
100%


Steps to Reproduce:
1. Configure network up.
2. Configure netconsole.
3. Try change network, such change network from dhcp to static or configure another nic.
 
Actual results:
Kernel panic happen "Kernel panic - not syncing: Watchdog detected hard
LOCKUP on cpu 1".

Expected results:
No kernel panic happen.

Additional info:
Comment 2 Mike Burns 2012-05-25 11:05:35 UTC 
Is this reproducible on RHEL?
Comment 3 Guohua Ouyang 2012-05-25 11:37:44 UTC 
Created attachment 586835 [details]
kernel_panic screenshot

(In reply to comment #2)
> Is this reproducible on RHEL?

attach the screenshot firstly.

test on RHEL63 2.6.32-272.el6:
1. configure network with dhcp
2. loaded netconsole.
3. change network from dhcp to static.
4. service network restart

no kernel panic happened.
Comment 4 Mike Burns 2012-05-29 12:22:56 UTC 
Any additional info on the panic?  Stack trace? core dump?  anything?
Comment 5 Fabian Deutsch 2012-05-29 16:06:14 UTC 
I could not reproduce this problem with snap 5 using the followign steps:

1. Fresh install
2. Configure network up (dhcp)
3. Configure netconsole (server set to 127.0.0.1)
4. Reconfigure network to static

No kernel oops appeared (nothing on the console or dmesg)

Can you give any more details, Ouyangguohua?
Comment 8 Guohua Ouyang 2012-05-30 02:11:45 UTC 
(In reply to comment #5)
> I could not reproduce this problem with snap 5 using the followign steps:
> 
> 1. Fresh install
> 2. Configure network up (dhcp)
> 3. Configure netconsole (server set to 127.0.0.1)
> 4. Reconfigure network to static
> 
> No kernel oops appeared (nothing on the console or dmesg)
> 
> Can you give any more details, Ouyangguohua?

if you specify the 127.0.0.1 as the netconsole server, does the service netconsole is loaded? I tried it here the service is not loaded. 

I think you should specify another valid server to reproduce this bug.
Comment 9 Guohua Ouyang 2012-05-30 02:17:49 UTC 
(In reply to comment #4)
> Any additional info on the panic?  Stack trace? core dump?  anything?

The only thing can get is on the screen can read "Kernel panic - not syncing: Watchdog detected hard LOCKUP on cpu 6".  the screenshot is attaching in comment #3.

The screen is hang when pop-up this message, send "CAD" does not work.
Comment 10 Fabian Deutsch 2012-05-30 07:31:16 UTC 
(In reply to comment #8)
> (In reply to comment #5)
> > I could not reproduce this problem with snap 5 using the followign steps:
> > 
> > 1. Fresh install
> > 2. Configure network up (dhcp)
> > 3. Configure netconsole (server set to 127.0.0.1)
> > 4. Reconfigure network to static
> > 
> > No kernel oops appeared (nothing on the console or dmesg)
> > 
> > Can you give any more details, Ouyangguohua?
> 
> if you specify the 127.0.0.1 as the netconsole server, does the service
> netconsole is loaded? I tried it here the service is not loaded. 

Now I was able to reproduce this bug:
1. Fresh install
2. Configure network (dhcp)
3. Configure netconsole (server 10.0.0.1)
4. Reconfigure network to static

You should open a second console before running the test so you can look at the trace using dmesg (run dmesg once before you raise the bug to get it into memory)

A second thing I noted was that the netconsole service seems not to be active on default: chkconfig --list netconsole only shows off
Comment 11 Fabian Deutsch 2012-05-30 07:31:57 UTC 
Created attachment 587619 [details]
A trace
Comment 13 cshao 2012-06-08 03:07:06 UTC 
Created attachment 590329 [details]
netconsole kernel panic
Comment 14 Guohua Ouyang 2012-06-08 03:13:54 UTC 
Hi mburns,

According to comment #12, restart network service will kernel panic after configure netconsole.  This blocks registering to rhevm also if the netconsole is configured. should we bring the bug back to 6.3 or need a technical note?

Thanks,
Comment 15 cshao 2012-06-08 05:09:05 UTC 
Test in RHEL6.3-20120606.3
kernel-2.6.32.278.el6.x86_64

Test result:
Restart network service will hung on "Shutting down interface eth0" after configure netconsole.

==========================
# cat /etc/issue
Red Hat Enterprise Linux Server release 6.3 (Santiago)
Kernel \r on an \m
# uname -a
Linux hp-slot6-eth0.qe.lab.eng.nay.redhat.com 2.6.32-278.el6.x86_64 #1 SMP Fri Jun 1 14:17:12 EDT 2012 x86_64 x86_64 x86_64 GNU/Linux
# /etc/init.d/netconsole status
netconsole module loaded
# /etc/init.d/network restart
Shutting down interface br0:  [ok]
Shutting down interface eth0:
Comment 18 Fabian Deutsch 2012-07-11 12:13:06 UTC 
I can also reproduce this on RHEL 6.3 with kernel 2.6.32-279 :

1. Install rhel
2. Configure netconsole to point to 10.42.0.1
3. service netconsole start
4. service network restart

And fails to stop the network interfaces (like in bug #816034)
Comment 26 errata-xmlrpc 2013-02-28 16:34:30 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHBA-2013-0556.html