Bug 15380

Summary: KDM crashes with kernel-3.19.0-desktop-{3,4,5,6}.mga5 (I/O errors)
Product: Mageia Reporter: Barry Jackson <zen25000>
Component: RPM PackagesAssignee: Thomas Backlund <tmb>
Status: RESOLVED INVALID QA Contact:
Severity: major    
Priority: Normal CC: thierry.vignaud, tmb
Version: Cauldron   
Target Milestone: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Source RPM: kernel CVE:
Status comment:
Attachments: output of hwinfo
screenshot after waiting a few minutes
screenshot after waiting longer

Description Barry Jackson 2015-02-27 15:17:06 CET
Description of problem:
When using kernel-3.19.0-desktop-3.mga5, KDM crashes (total machine lock with no keyboard access to ttys - only option to power off.
However if the machine is left for several minutes after the 'lock-up' then a tty appears with apparent errors that continue ad infinitum - some blurry screen shots of these will be attached below along with the hwinfo output from the machine when running kernel-3.19.0-desktop-3.mga5.

Starting into CLI and running startx with the same kernel works fine.


Booting the previous kernel-3.19.0-desktop-2.mga5 works fine. with KDM.

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


How reproducible:


Steps to Reproduce:
1. Update to kernel-3.19.0-desktop-3.mga5 and reboot.
2.
3.


Reproducible: 

Steps to Reproduce:
Comment 1 Barry Jackson 2015-02-27 15:24:54 CET
Created attachment 5955 [details]
output of hwinfo
Comment 2 Barry Jackson 2015-02-27 15:26:13 CET
Created attachment 5956 [details]
screenshot after waiting a few minutes
Comment 3 Barry Jackson 2015-02-27 15:26:58 CET
Created attachment 5957 [details]
screenshot after waiting longer
Comment 4 Barry Jackson 2015-02-27 18:41:35 CET
Just updated to kernel-3.19.0-desktop-4.mga5 - no change.

However testing my second Cauldron fully updated today I don't see this issue, so am running e2fsck -fcky on all partitions used by my primary Cauldron with this problem.
Still seems odd that it is fine with kernel-3.19.0-desktop-2.mga5 and earlier if it's a filesystem issue.
Not sure if it may be relevant but / is on an SSD.
Comment 5 Barry Jackson 2015-02-27 21:00:24 CET
After checking the fs (nothing major found) I managed to get past KDE login - just long enough to switch to tty12 and see the same errors as before, switch back to tty1 in time to see the GUI crash.

Rebooted without GUI and used startx and no errors in tty12 and all is working fine :\

Any ideas Thomas?

CC: (none) => tmb

Thierry Vignaud 2015-02-28 06:49:50 CET

CC: (none) => thierry.vignaud
Summary: KDM crashes with kernel-3.19.0-desktop-3.mga5 => KDM crashes with kernel-3.19.0-desktop-3.mga5 (I/O errors)
Source RPM: (none) => kernel

Comment 6 Barry Jackson 2015-03-02 20:34:56 CET
Just updated to kernel-3.19.0-desktop-5.mga5 - no change.

As well as the I/O errors in tty12 I noticed mentions of 'Read only file system' when mysqld was attempting to start and failing.

CLI login and startx still works fine (I'm using kernel-3.19.0-desktop-5.mga5 now)

Priority: Normal => High

Comment 7 Barry Jackson 2015-03-08 22:23:37 CET
kernel-3.19.0-desktop-6.mga5 - no change.

No one else seems to be affected by this and it only affects my main system. My second Cauldron fallback install (using the same /home) does not exhibit this, so I am switching to this for daily use. 

Maybe something will show up as I add some of the applications not yet included in this installation.
I will keep the original for now in case any tests are needed.

Dropping priority to normal and severity to major.

Priority: High => Normal
Severity: critical => major

Barry Jackson 2015-03-08 22:24:20 CET

Summary: KDM crashes with kernel-3.19.0-desktop-3.mga5 (I/O errors) => KDM crashes with kernel-3.19.0-desktop-{3,4,5,6}.mga5 (I/O errors)

Samuel Verschelde 2015-05-21 12:41:14 CEST

Assignee: bugsquad => tmb

Samuel Verschelde 2015-06-06 16:20:04 CEST

Whiteboard: (none) => MGA5TOO

Comment 8 Thomas Backlund 2015-09-18 15:40:23 CEST
is this still an issue ?

kernel-side there is now 4.1.7 in cauldron and mga5 testing
Comment 9 Barry Jackson 2015-09-19 12:53:52 CEST
Thomas - sorry I forgot all about this.

The I/O errors were eventually traced to a failing SSD, which I replaced.

Since then I have not seen this issue.

So closing as invalid.

Status: NEW => RESOLVED
Resolution: (none) => INVALID
Whiteboard: MGA5TOO => (none)