We have 2 machines (quad core intel i5) running suse 11.2. They were clean installs and both suffer from this problem on around 50% of boot-ups. Other times, the system boots quickly and is fine.
Basically, one of the CPU’s gets hammered to 100% (according to KDE system monitor) for around 10 minutes after boot up. Although the other three CPU’s seem mostly idle, the system is very slow, to the point of being unusable until suddenly the system recovers and runs normally.
I’ve looked at ‘top’ and the KDE system monitor and both show no process taking more than a few % of the CPU. So it is a mystery as to what is taking up so much CPU and why it does it some days and not others !
One other thing, if you try to run virtualbox during this time, it (eventually) says that the kernel drives are not loaded - so possibly the kernel is stuck loading drivers. Infact, from dmesg, I can see that the system is still booting but other than the extended time stamps, the only obvious difference between a good boot and a bad one seems to be the line :
141.794727] hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj.
which is there after a slow boot. The sound works ok (as does everything else).
Does anyone have any ideas where I should look from here ?
Are the machines being shut down properly? Are you seeing lots of disk activity? This appears to be happening after you log in and the GUI is running, so the boot process should be done at that point.
since both systems display the same unusual symptoms, maybe you have
a faulty install from a corrupted install image?
- get your install image from http://software.opensuse.org/112/en ?
(if not, then where?)
- check the md5sum of the downloaded iso prior to burning the disk?
- do this http://tinyurl.com/yajm2aq before install attempt?
if you answered “no” (or “don’t know”) to any of those then see the
following cites before you start over:
but, if your install image checks out ok
and you had no errors during install
and you have not enabled more than the basic four repos (oss,
non-oss, update and packman)
and you have updated your machine correctly using YaST or the YaST
then you need to post again and tell us you had a good image, etc etc
etc…and, tell us: have both machines had these same symptoms from
the very first boot…or, was it all smooth for a while and then blam,
DenverD (Linux Counter 282315)
posted via NNTP w/TBird 184.108.40.206 | KDE 3.5.7 | openSUSE 10.3
220.127.116.11-0.4-default SMP i686
AMD Athlon 1 GB RAM | GeForce FX 5500 | ASRock K8Upgrade-760GX |
CMedia 9761 AC’97 Audio
Well, this is embarrassing, this morning I can’t get either machine to do this, no matter how many reboots I do. :\
Both machines always shut-down cleanly. If / when it happens again I’ll check the disk activity.
I did get the install disk image from the suse link you specified but did not bother checking it as described. Also, we are using repos other than those listed (eg nvidia, virtualbox and a few others) so I’ll look into doing a re-install from a verified image if the problem re-occurs.
We didn’t notice the machines doing this at first, so probably it is something we installed. What is strange is that I don’t know how the CPU usage can be high but no process is actually using any CPU. It seems like there must be some other processes going on that don’t get listed by ‘top’ and I was hoping there might be a way to see what the CPU is actually running, rather than to have to resort to removing applications until it goes away.
Anyway, thanks very much for your help,
Did some searching. First thing I see, is that the message concerns the sound card. Please search the forums for ‘position_fix’ and ‘bdl_pos_adj’, there are some posts regarding this matter. The sound card could very well be the cause of the slow booting. It looks like the kernel is coming into some loop during boot.
In case anyone else has this problem (which persisted on and off even with upgrade to 11.3), the issue was that the hard disk was being read constantly at a very high rate.
Un-installing ‘preload’ fixed the problem for us.