Bug 590923

Summary: /sbin/mkdumprd spins forever doing "readlink"
Product: [Fedora] Fedora Reporter: Roland McGrath <roland>
Component: kexec-toolsAssignee: Neil Horman <nhorman>
Status: CLOSED ERRATA QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: medium Docs Contact:
Priority: low    
Version: 13CC: nhorman
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: kexec-tools-2.0.0-35.fc13 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2010-05-13 19:26:32 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:
Attachments:
Description Flags
-x log
none
patch to fix how we sift through sysfs
none
updated patch none

Description Roland McGrath 2010-05-11 00:12:28 UTC 
Description of problem:

First noticed during boot with kexec-tools installed,
also happens with manual "service kexec-tools start".
It spins forever chewing CPU.
I found that running /sbin/mkdumprd is where it happens and doing this by hand hits it too.

I've attached the log of running "bash -x /sbin/mkdumprd ..." where it eventually runs "readlink /sys/class/block/sda5" repeatedly forever.

Version-Release number of selected component (if applicable):
kexec-tools-2.0.0-34.fc13.x86_64

How reproducible:


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


Expected results:


Additional info:
Comment 1 Neil Horman 2010-05-11 11:10:33 UTC 
could you attach that file please?  thanks !
Comment 2 Roland McGrath 2010-05-11 16:24:25 UTC 
Created attachment 413187 [details]
-x log

Oops, thought I'd attached it in the initial report.
Comment 3 Neil Horman 2010-05-12 15:40:26 UTC 
Created attachment 413462 [details]
patch to fix how we sift through sysfs

looks like I forgot to port a RHEL patch over to Fedora.  Let me know if this fixes your problem.  Thanks!
Comment 4 Roland McGrath 2010-05-12 18:11:39 UTC 
With that, the infinite loop is instead:

++ readlink ./device
Comment 5 Neil Horman 2010-05-12 18:46:55 UTC 
Created attachment 413521 [details]
updated patch

sorry, missed a hunk from the backport.
Comment 6 Roland McGrath 2010-05-12 19:26:33 UTC 
That fixes mkdumprd I guess.  It completed.  But "service kdump start" is not happy.  It just says "FAILED" with no details.  syslog gets:

May 12 12:25:46 magilla kdump: No crashkernel parameter specified for running kernel
May 12 12:25:46 magilla kdump: failed to start up
Comment 7 Neil Horman 2010-05-12 19:28:32 UTC 
It told you exactly what the problem was:

"No crashkernel parameter specified for running
kernel"

you need to specify a crashkernel command line parameter on the kernel command line.  For x86 system crashkernel=128M is usually sufficient.

I'll commit this change shortly.  Thanks!
Comment 8 Roland McGrath 2010-05-12 19:37:47 UTC 
I'm using what you get out of the box for booting F-13 with kexec-tools installed.  I don't want to fiddle kernel parameters by hand.  Either it should just work, or it should not be chkconfig on by default.
Comment 9 Fedora Update System 2010-05-12 20:10:54 UTC 
kexec-tools-2.0.0-35.fc13 has been submitted as an update for Fedora 13.
http://admin.fedoraproject.org/updates/kexec-tools-2.0.0-35.fc13
Comment 10 Fedora Update System 2010-05-13 19:26:27 UTC 
kexec-tools-2.0.0-35.fc13 has been pushed to the Fedora 13 stable repository.  If problems still persist, please make note of it in this bug report.