Bug 1044361

Summary: systemptap script seldomly caused kernel panic on redhat 5
Product: Red Hat Enterprise Linux 5 Reporter: Lu Wang <wanglu621>
Component: systemtapAssignee: Frank Ch. Eigler <fche>
Status: CLOSED INSUFFICIENT_DATA QA Contact: qe-baseos-tools-bugs
Severity: high Docs Contact:
Priority: unspecified    
Version: 5.5CC: mbenitez, mjw, scox, wanglu621
Target Milestone: rcFlags: fche: needinfo? (wanglu621)
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-02-11 16:13:45 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:

Description Lu Wang 2013-12-18 08:32:45 UTC 
Description of problem:
Hi, 
A systemtap module which records file system IO pattern(distribution of read,write, read seek, write seek and  R/W throughput) has caused several times of kernel panics on our computing cluster. 

Version-Release number of selected component (if applicable):
systemtap-1.6-7.el5_8.x86_64
systemtap-runtime-1.6-7.el5_8.x86_64

How reproducible:

30 kernel panics / 20000+ successful runs 

30 nodes/ 900 nodes 


Steps to Reproduce:
1. run the script with a cron job which finds unprobed jobs on the computing node every 15 mins.
2.
3.

Actual results:

some computing nodes crashed

Expected results:


Additional info:
The script was compiled with command "stap -r "2.6.18-348.3.1.el5"  -DMAXMAPENTRIES=10240 client.stp -m iopattern.ko
" and run by command "staprun -o myoutput.raw -x pid iopattern.ko"

Unfortunately the client node has not configured crashdump, the last screen output is also attached. 

An normal output of the script is also attached.
Comment 1 Frank Ch. Eigler 2013-12-18 12:30:24 UTC 
Please attach the last-screen-output and other basic content if able,
plus the output of "stap-report".  Depending on the details, it may
suggest a problem in systemtap, or the kernel.

Please note that a slightly newer version 1.8 of systemtap was included
in a later verion of RHEL5, and a newer one yet in the Developer Toolset
package.  Plus one can build a version out of upstream/git which will
work on RHEL5.
Comment 2 Frank Ch. Eigler 2014-01-28 16:09:40 UTC 
Please note that without more information, we will be unable to diagnose the problem further.  If this problem still occurs, please supply data as per comment #1.