Bug 445080

Summary:	named kills machine
Product:	[Fedora] Fedora	Reporter:	Neal Becker <ndbecker2>
Component:	bind	Assignee:	Adam Tkac <atkac>
Status:	CLOSED DUPLICATE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	low	Docs Contact:
Priority:	low
Version:	rawhide	CC:	atkac, ovasik
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:	2008-05-05 20:39:27 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 Neal Becker 2008-05-03 11:24:04 UTC

Description of problem:

I just tried a 2nd upgrade F8-> rawhide.

On reboot, it hung a long time on bringing up eth0, and failed.  This never
had a problem on F8.

06:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B
PCI Express Gigabit Ethernet controller (rev 01)

When it went to start X, all I got was an 'X'.  And an apparantly dead
machine.

I rebooted into mode 3.  Ran top.  Nothing showed a significant load
average.  Quit top.  Now, no response from keyboard.

Rebooted into mode 1.  Looked at /var/log/messages.  Millions(?) of lines
from named.  chkconfig named off.  Rebooted again.  Now, running OK, but
still no eth0.

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


How reproducible:


Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:

Comment 1 Adam Tkac 2008-05-05 07:22:59 UTC

Named cannot break network device configuration.
- what kind of messages are in system log?
- are you using NetworkManager?
- are you sure this is not SELinux problem?

Thanks, Adam

Comment 2 Neal Becker 2008-05-05 09:52:32 UTC

Sorry for not being clear.  named did not break network config.  What happened
was:

1. eth0 was not brought up (for some other reason)
2. named went nuts, logging thousands of error messages, apparantly generating
so much load that machine was unusable.  (This diagnosis is a guess, but see
#3)
3. turning off named fixed the problem
4. selinux is/was disabled

Comment 3 Adam Tkac 2008-05-05 12:12:20 UTC

Would it be possible attach what messages were in log, please? I'm not able
track what happen without them.

Comment 4 Neal Becker 2008-05-05 12:31:00 UTC

They all look like this:

May  2 05:53:53 nbecker1 named[2410]: too many timeouts resolving
'localhost.md.hnsnet/AAAA' (in '.'?): disabling EDNS
May  2 05:53:54 nbecker1 named[2410]:last message repeated 3 times
May  2 05:53:54 nbecker1 named[2410]: too many timeouts resolving 'A.ROOT-
May  2 05:53:54 nbecker1 named[2410]: too many timeouts resolving
'localhost.md.hnsnet/AAAA' (in '.'?): disabling EDNS
May  2 05:53:59 nbecker1 named[2410]:last message repeated 3 times
May  2 05:53:54 nbecker1 named[2410]: too many timeouts resolving 'B.ROOT-
May  2 05:54:00 nbecker1 named[2410]:last message repeated 4 times
May  2 05:53:54 nbecker1 named[2410]: too many timeouts resolving 'C.ROOT-

Comment 5 Adam Tkac 2008-05-05 20:39:27 UTC

hm, I don't think those messages can kill machine. But bug about this "issue" is
already opened...

*** This bug has been marked as a duplicate of 275091 ***