Bug 694214 - RHELS 5.5 running in kvm: sosreport tries to read ide smart values and generates kernel errors
Summary: RHELS 5.5 running in kvm: sosreport tries to read ide smart values and genera...
Keywords:
Status: CLOSED NEXTRELEASE
Alias: None
Product: Red Hat Enterprise Linux 5
Classification: Red Hat
Component: kernel
Version: 5.5
Hardware: x86_64
OS: Linux
unspecified
low
Target Milestone: rc
: ---
Assignee: Virtualization Maintenance
QA Contact: Red Hat Kernel QE team
URL:
Whiteboard:
: 860796 (view as bug list)
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-04-06 18:21 UTC by Pierre Carrier
Modified: 2012-10-04 08:28 UTC (History)
9 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of: 678261
Environment:
Last Closed: 2012-06-07 07:51:30 UTC
Target Upstream Version:


Attachments (Terms of Use)

Description Pierre Carrier 2011-04-06 18:21:59 UTC
+++ This bug was initially created as a clone of Bug #678261 +++

Description of problem:

sosreport copies smart_values and smart_thresholds from /proc/ide/ide?/hd?

Doing this in kvm virtual machine generates ide errors:

Feb 17 11:06:52 <kern.warn> HOSTNAME kernel: ide: failed opcode was: 0xb0
Feb 17 11:06:52 <kern.warn> HOSTNAME kernel: hda: task_in_intr: status=0x41 { DriveReady Error }
Feb 17 11:06:52 <kern.warn> HOSTNAME kernel: hda: task_in_intr: error=0x04 { DriveStatusError }
Feb 17 11:06:52 <kern.warn> HOSTNAME kernel: ide: failed opcode was: 0xb0
Feb 17 11:23:49 <kern.warn> HOSTNAME kernel: hda: task_no_data_intr: status=0x41 { DriveReady Error }
Feb 17 11:23:49 <kern.warn> HOSTNAME kernel: hda: task_no_data_intr: error=0x04 { DriveStatusError }
Feb 17 11:23:49 <kern.warn> HOSTNAME kernel: ide: failed opcode was: 0xb0
Feb 17 11:23:49 <kern.warn> HOSTNAME kernel: hda: task_in_intr: status=0x41 { DriveReady Error }
Feb 17 11:23:49 <kern.warn> HOSTNAME kernel: hda: task_in_intr: error=0x04 { DriveStatusError }
Feb 17 11:23:49 <kern.warn> HOSTNAME kernel: ide: failed opcode was: 0xb0

Version-Release number of selected component (if applicable): sos-1.7-9.49.el5


How reproducible:
Run sosreport in a kvm virtual machine.  Or just "cat /proc/ide/ide?/hd?/smart_values".

Steps to Reproduce:
1. Run sosreport in virtual machine
2. look for ide errors
3.
  
Actual results:
ide errors are reported when sosreport runs

Expected results:
sosreport should not cause any error messages to be reported.

Additional info:

--- Additional comment from prc on 2011-02-23 18:50:39 GMT ---

Hello Andreas,


Thank you for your feedback.

The RHEL5 IDE driver warns you (syslog level kernel.warn) that it cannot get the SMART status of the virtual IDE drive.

What would you like to change? Should kvm answer in a nicer way? Should the kernel log message "look less harmful"?

Either way, that's not a bug in sosreport, we try to get this info on purpose.
Depending on your expectations, I believe this is either CLOSE DUPLICATE of 676264/663665, or reassigned to the appropriate component.


Yours sincerely,

-- 
Pierre Carrier, Technical Support Engineer
Production Support, EMEA office
Global Support Services
Red Hat, Inc.

--- Additional comment from andreas on 2011-02-23 20:08:27 GMT ---

(In reply to comment #1)
Hi Pierre,

we have a number of production servers that run sosreport from cron every day.  Also we run a log reporting tool to analyze the (messages) log files.  On kvm machines we get a daily report with this kernel warning.

Ideally we would like to stop unnecessary warnings and errors from being written to the logs in the first place - before we then filter with log reporting.

It would be nice if either the kernel ide driver in kvm would not offer the non-retrievable smart statistics, or if sosreport was made aware that getting smart values from Qemu virtual disks is useless and should not be done.  (just like it is best practice to disable smartd in virtual machines)

In short: it would be nice if sosreport could run in kvm without generating any warnings at all.

I understand that the logged messages are not errors but a specific kernel warning about smart values being not retrievable from the disks.  So for our specific setup it would be okay to just filter them out.  However, the drive ready and drive status errors might be reported for other reasons as well.

So another solution would be to make the errors nicer so we can ignore them for our log reports.  Just as you suggested.

Thanks for your time.

Best regards,

Andreas Obermaier

Comment 1 Jes Sorensen 2012-04-30 20:11:42 UTC
Looks like issue with IDE emulation in qemu-kvm - reassigning to virt team.

Comment 9 Ronen Hod 2012-10-04 08:28:05 UTC
*** Bug 860796 has been marked as a duplicate of this bug. ***


Note You need to log in before you can comment on or make changes to this bug.