Bug 620474

Summary:	kdump service fails to start
Product:	Red Hat Enterprise Linux 6	Reporter:	michal novacek <mnovacek>
Component:	kexec-tools	Assignee:	Cong Wang <amwang>
Status:	CLOSED NOTABUG	QA Contact:	Red Hat Kernel QE team <kernel-qe>
Severity:	medium	Docs Contact:
Priority:	low
Version:	6.0	CC:	rkhan, syeghiay
Target Milestone:	rc	Keywords:	Reopened, RHELNAK
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2010-08-20 14:17:47 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 michal novacek 2010-08-02 15:54:33 UTC

Description of problem: 
kdump service will not start.


Version-Release number of selected component (if applicable):
rhel6, kernel 2.6.32-44.el6.x86_64
kexec-tools-2.0.0-94.el6.x86_64

How reproducible: always

Steps to Reproduce:
1. boot rhel6 with added "crashkernel=128M@16M" parameter to kernel command line
2. run: /etc/init.d/kdump restart

  
Actual results: failure of the script

Expected results: success. 

Additional info:
my guess is that there is a change in the way how to know whether the parameter was specified on kernel command line.

# uname -r
2.6.32-44.el6.x86_64

# cat /etc/redhat-release 
Red Hat Enterprise Linux Server release 6.0 Beta (Santiago)

# cat /proc/cmdline 
ro root=/dev/mapper/vg_dellpe86001-lv_root rd_LVM_LV=vg_dellpe86001/lv_root rd_LVM_LV=vg_dellpe86001/lv_swap rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=us console=ttyS0,115200 crashkernel=128M@16M

# /etc/init.d/kdump restart
Memory for crashkernel is not reserved
Please reserve memory by passing "crashkernel=X@Y" parameter to the kernel
Stopping kdump:                                            [FAILED]
Starting kdump:                                            [FAILED]

# tail /var/log/messages
Aug  2 11:41:42 dell-pe860-01 init: prefdm main process ended, respawning
Aug  2 11:41:42 dell-pe860-01 init: prefdm main process (2057) terminated with status 1
Aug  2 11:41:42 dell-pe860-01 init: prefdm main process ended, respawning
Aug  2 11:41:42 dell-pe860-01 init: prefdm main process (2074) terminated with status 1
Aug  2 11:41:42 dell-pe860-01 init: prefdm respawning too fast, stopped
Aug  2 11:41:42 dell-pe860-01 ntpd[1682]: Listening on interface #7 virbr0, 192.168.122.1#123 Enabled
Aug  2 11:42:44 dell-pe860-01 kdump: kexec: failed to unload kdump kernel
Aug  2 11:42:44 dell-pe860-01 kdump: failed to stop
Aug  2 11:42:44 dell-pe860-01 kdump: No crashkernel parameter specified for running kernel
Aug  2 11:42:44 dell-pe860-01 kdump: failed to start up

# cat /sys/kernel/kexec_crash_size
0

Comment 2 RHEL Program Management 2010-08-02 16:27:36 UTC

This issue has been proposed when we are only considering blocker
issues in the current Red Hat Enterprise Linux release.

** If you would still like this issue considered for the current
release, ask your support representative to file as a blocker on
your behalf. Otherwise ask that it be considered for the next
Red Hat Enterprise Linux release. **

Comment 3 Cong Wang 2010-08-03 01:24:33 UTC

This is not a bug, your kernel failed to find a 128M hole at 16M, thus no memory is reserved. You should try to use 128M@0 to let the kernel to find a base address.

Comment 4 michal novacek 2010-08-19 15:00:15 UTC

Works well. Two questions though before you close it:
 1] is kdump script supposed to work with "crashkernel=auto" parameter? If it is it does not, saying that you have not passed "crashkernel=X@Y".
 2] how do I tell correct value? the only source of information I found is the /usr/share/doc/*/ howto that does not give the explanation. it only tells you that you should give "128M@16M" for i386 and x86_64.

Comment 5 Cong Wang 2010-08-20 02:40:22 UTC

(In reply to comment #4)
> Works well. Two questions though before you close it:
>  1] is kdump script supposed to work with "crashkernel=auto" parameter? If it
> is it does not, saying that you have not passed "crashkernel=X@Y".


kdump will not see "crashkernel=auto", it will see the actual number in /proc/cmdline.

>  2] how do I tell correct value? the only source of information I found is the
> /usr/share/doc/*/ howto that does not give the explanation. it only tells you
> that you should give "128M@16M" for i386 and x86_64.

There is no one true way, this is why "crashkernel=auto" gets introduced. Usually you can try 128M@0 or 256M@0 for x86 machines.