Bug 661894 - Buggy firmware of SSD not detected (was: /sbin/runlevel reports 5, but no services from runlevel 5 have been started)
Summary: Buggy firmware of SSD not detected (was: /sbin/runlevel reports 5, but no ser...
Status: RESOLVED WORKSFORME
Alias: None
Product: openSUSE 11.3
Classification: openSUSE
Component: Kernel (show other bugs)
Version: Final
Hardware: x86-64 openSUSE 11.3
: P5 - None : Normal with 1 vote (vote)
Target Milestone: ---
Assignee: E-mail List
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-12-30 18:36 UTC by Martin Schröder
Modified: 2011-05-04 13:54 UTC (History)
3 users (show)

See Also:
Found By: Community User
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
dmesg (44.29 KB, text/plain)
2010-12-30 18:37 UTC, Martin Schröder
Details
hwinfo --all (40.15 KB, text/plain)
2010-12-30 18:38 UTC, Martin Schröder
Details
boot.msg (54.25 KB, text/plain)
2011-02-23 20:18 UTC, Martin Schröder
Details
/var/log/messages from boot (15.34 KB, text/plain)
2011-02-23 20:22 UTC, Martin Schröder
Details
rpm -aV (2.25 KB, text/plain)
2011-02-25 19:20 UTC, Martin Schröder
Details
ls -laR /etc/init.d/ (23.15 KB, text/plain)
2011-02-25 19:21 UTC, Martin Schröder
Details
.depend.start (1.44 KB, text/plain)
2011-02-25 19:22 UTC, Martin Schröder
Details
insserv -dv (2.51 KB, text/plain)
2011-02-25 19:25 UTC, Martin Schröder
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Schröder 2010-12-30 18:36:33 UTC
User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; de; rv:1.9.2.13) Gecko/20101203 SUSE/3.6.13-3.1 Firefox/3.6.13

After a boot, /sbin/runlevel reports "N 5" (which is indeed the default runlevel), but _no_ services from level 5 have been started; they must all be started manually.


Reproducible: Always

Steps to Reproduce:
1. Reboot
Actual Results:  
System reports runlevel 5, but no services from runlevel 5 have been started.

Expected Results:  
All services from level 5 should be started.
Comment 1 Martin Schröder 2010-12-30 18:37:27 UTC
Created attachment 406613 [details]
dmesg
Comment 2 Martin Schröder 2010-12-30 18:38:33 UTC
Created attachment 406614 [details]
hwinfo --all
Comment 3 Martin Schröder 2011-01-03 23:48:33 UTC
No change with 2.6.34.7-0.7, although I was greeted by kdm after rebooting - but only once (and I didn't check if other services from level 5 were running); several reboots only provided a non-functioning level 5.
An "init 3;init 5" started a kdm, but most other services from level 5 were missing.
Comment 4 Dr. Werner Fink 2011-01-10 08:49:27 UTC
can not reproduce that is it works as expected ... beside this the only
service between runlevel 3 and runlevel 5 is normally the display manager
aka kdm.  There is no need to start services which are already running.
A runlevel switch only stops services which are not enabled in the new
runlevel and only starts services which are only enabled in the new runlevel.
Comment 5 Martin Schröder 2011-02-23 19:58:55 UTC
(In reply to comment #4)
> can not reproduce that is it works as expected ... beside this the only
> service between runlevel 3 and runlevel 5 is normally the display manager
> aka kdm.  There is no need to start services which are already running.

Did you even read what I reported? This is not about a switch but about BOOT.

No services from /etc/init.d/rc5.d have been started automatically. runlevel3 is of no relevance, as the system boots directly into 5. 

I understand that it normally works as intended, but I'm baffled at what's the cause for this problem.
Comment 6 Martin Schröder 2011-02-23 20:18:16 UTC
Created attachment 415847 [details]
boot.msg
Comment 7 Martin Schröder 2011-02-23 20:22:24 UTC
Created attachment 415848 [details]
/var/log/messages from boot
Comment 8 Dr. Werner Fink 2011-02-25 08:49:15 UTC
(In reply to comment #5)

Report

  rpm -aV
  ls -l /etc/init.d/rc5.d/
  cat /etc/init.d/.depend.start

... IMHO you touched the boot system without any need, to restore
youre system you may run

  insserv -d

Beside this bugzilla is not for repairing system broken by the users.
Comment 9 Martin Schröder 2011-02-25 19:20:15 UTC
Created attachment 416335 [details]
rpm -aV
Comment 10 Martin Schröder 2011-02-25 19:21:32 UTC
Created attachment 416336 [details]
ls -laR /etc/init.d/
Comment 11 Martin Schröder 2011-02-25 19:22:51 UTC
Created attachment 416337 [details]
.depend.start
Comment 12 Martin Schröder 2011-02-25 19:25:11 UTC
(In reply to comment #8)
> (In reply to comment #5)
> Report
> 
>   rpm -aV
>   ls -l /etc/init.d/rc5.d/
>   cat /etc/init.d/.depend.start

All attached.

> ... IMHO you touched the boot system without any need, to restore
> youre system you may run
> 
>   insserv -d

That report is also attached; it didn't help.

> Beside this bugzilla is not for repairing system broken by the users.

I'm not aware of having broken this system.
Comment 13 Martin Schröder 2011-02-25 19:25:45 UTC
Created attachment 416338 [details]
insserv -dv
Comment 14 Martin Schröder 2011-02-27 15:12:19 UTC
The boot device of the system is an intel ssd:
----------------
15: IDE 00.0: 10600 Disk
  [Created at block.245]
  Unique ID: 3OOL.tdobUQ+gv7C
  Parent ID: w7Y8.VAeuMo4mTH6
  SysFS ID: /class/block/sda
  SysFS BusID: 0:0:0:0
  SysFS Device Link: /devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0
  Hardware Class: disk
  Model: "INTEL SSDSA2M080"
  Vendor: "INTEL"
  Device: "SSDSA2M080"
  Revision: "2CV1"
  Serial ID: "CVPO017402JP080JGN"
  Driver: "ahci", "sd"
  Driver Modules: "ahci"
  Device File: /dev/sda
  Device Files: /dev/sda, /dev/block/8:0, /dev/disk/by-id/ata-INTEL_SSDSA2M080G2GN_CVPO017402JP080JGN, /dev/disk/by-id/scsi-SATA_INTEL_SSDSA2M08CVPO017402JP080JGN, /dev/disk/by-path/pci-0000:00:1f.2-scsi-0:0:0:0, /dev/disk/by-id/wwn-0x5001517959331252
  Device Number: block 8:0-8:15
  BIOS id: 0x80
  Geometry (Logical): CHS 9729/255/63
  Size: 156301488 sectors a 512 bytes
  Geometry (BIOS EDD): CHS 155061/16/63
  Size (BIOS EDD): 156301488 sectors
  Geometry (BIOS Legacy): CHS 1023/255/63
  Config Status: cfg=no, avail=yes, need=no, active=unknown
  Attached to: #11 (SATA controller)
----------------
I updated its firmware and was now greeted by kdm after boot,
so everything seems to be fine now. I'm still confused how a broken
firmware could cause this...
Comment 15 Dr. Werner Fink 2011-02-28 09:38:44 UTC
Strange ... the only problem could be that the preload code
could cause problems on an ssd, for an ssd I would disable
preload by

       rpm -e preload --force --nodeps

nevertheless IMHO there was a bug in the firmware of the ssd
which has caused a race e.g. wrong order of the dependencies
of the boot scripts. Or high I/O load has leaded to missed
reads from the ssd breaking scripts, programs, and/or libraries.
Comment 16 Martin Schröder 2011-02-28 10:02:36 UTC
(In reply to comment #15)
> nevertheless IMHO there was a bug in the firmware of the ssd
> which has caused a race e.g. wrong order of the dependencies
> of the boot scripts. Or high I/O load has leaded to missed
> reads from the ssd breaking scripts, programs, and/or libraries.

I can only say that it worked when I installed 11.3 on the system 
(with the old firmware) sometime in October 2010, and stopped 
working sometime later in 2010. Since I don't reboot everyday but 
use hibernate, I can't pinpoint the exact date.
Comment 17 Jeff Mahoney 2011-04-25 19:20:48 UTC
I'm not sure what we can do here. There's nothing in the log indicating a problem has occurred, either from the kernel or userspace.
Comment 18 Martin Schröder 2011-05-04 13:52:47 UTC
Just a heads up: The new kernel for 11.3 brought with it a reboot. 
Now I'm again greeted by a console login prompt and find no services running.
Maybe an upgrade to 11.4 will solve this...
Comment 19 Martin Schröder 2011-05-04 13:54:11 UTC
(In reply to comment #15)
> Strange ... the only problem could be that the preload code
> could cause problems on an ssd, for an ssd I would disable
> preload by
> 
>        rpm -e preload --force --nodeps

rpm: only installation, upgrading, rmsource and rmspec may be forced :-)

Should I delete preload-kmp-desktop also?