About an hour ago I resized my home partition, with gparted. All went well and I rebooted into Ubuntu Natty, no problem.
Shortly after I rebooted into suse 11.4 and checked a couple of forums and did one or two other minor jobs.
I then shutdown the laptop for about 20 minutes.
When I restarted it, choosing suse 11.4, it took about 3 minutes for all the text lines to complete. No error messages appeared of any kind.
The screen then went black, then the Nvidia splash screen appeared, then went black again (all normal) but it then took another 2 minutes to boot to the desktop.
All in all, it took more than 5 minutes to boot, with disk activity all the time. This laptop usually boots 11.4 in about 90 seconds, to a useable desktop.
When the desktop arrived it was virtually unuseable due to hard disk activity. I checked with System Monitor (which took about a minute to open) and it reported that 290MB of ram were in use (of 2GB), no processes were using any cpu except for system monitor (about 2%) but the hard disk activity light was solid. Response was very slow - as I say, virtually unuseable.
I REISUB’d and twice more it booted this way, and behaved the same way once booted. The last time, after about 5 minutes of (non) use, the hard disk avtivity light went out and everything returned to normal.
I have rebooted since and everything is back to normal.
Can anybody tell me what happened? Does anything run in 11.4 every now and again that must be completed, but which is not user requested (like some sort of system check)?
Thanks.
On 03/16/2011 03:36 PM, quackers wrote:
>
> Does anything run in 11.4 every now
> and again that must be completed, but which is not user requested (like
> some sort of system check)?
yep…
next time it happens, if it does instead of killing the machine with
all those REISUBs and reboots, try to learn what is going on…open a
terminal, type top and hit enter and watch what is sucking all
your CPU cycles…
look at top now to see if you can figure out most of what happens
there…if not, type man top and hit enter…
–
DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.1.8, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11
Thanks for your reply DenverD, though it seems somewhat self-contradicting!
Quote
> Does anything run in 11.4 every now
> and again that must be completed, but which is not user requested (like
> some sort of system check)?
yep…
next time it happens, if it does instead of killing the machine with
all those REISUBs and reboots, try to learn what is going on…open a
terminal, type top and hit enter and watch what is sucking all
your CPU cycles…
I ask if something runs every now and again, and you say “yep”, then say “next time it happens, if it does”.
Well you said it does!
What do you kill an unresponsive machine with, may I ask? The power button?
How do I “try to learn what’s going on” in a largely unresponsive machine? As I said in post 1, I waited for system monitor to load and it said virtually no cpu cycles were in use. Opening a terminal took ages too. Things taking ages was what I wanted to avoid.
I am also able to figure out “top” too. I have several conky displays, all of which show the 4 top, top processes.
Please don’t be condescending, as it’s not helpful.
It’s not safe to assume I know nothing, just because of a low post count.
On 2011-03-16 17:06, quackers wrote:
> How do I “try to learn what’s going on” in a largely unresponsive
> machine?
With patience 
Iotop, perhaps.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
And if it just continues without change? Just wait for the hard drive to die?
On 03/16/2011 05:06 PM, quackers wrote:
> How do I “try to learn what’s going on” in a largely unresponsive
> machine?
top or atop would be my first guesses…
> As I said in post 1, I waited for system monitor to load and it
> said virtually no cpu cycles were in use. Opening a terminal took ages
> too. Things taking ages was what I wanted to avoid.
well, wanting to avoid “taking ages” may also avoid learning what is
going on…
> I am also able to figure out “top” too. I have several conky displays,
> all of which show the 4 top, top processes.
i can’t see your experience from here, and you didn’t mention you have
already tried to learn what was going on with top, atop, htop or even
by looking in the logs…so, what am i supposed to do? assume you
have already done every think i can think of which might help–and not
try to help you??
> Please don’t be condescending, as it’s not helpful.
wow!
> It’s not safe to assume I know nothing, just because of a low post
> count.
i have NO idea what your post count is (i do not use the web forum and
therefore don’t know if you have 1 post of 100,000! so don’t assume i
have an agenda, OR think you are stupid)…and no matter how you
chose to read my attempt to help you it was meant to be helpful and
NOT meant to be condescending…
but yours was meant to be condescending, demeaning and hurtful, thanks…
i won’t be bothering you again.
–
DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.1.8, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11
We’ll agree to differ then.
On 2011-03-16 23:06, quackers wrote:
> And if it just continues without change? Just wait for the hard drive
> to die?
Are you suspecting the drive is bad?
If your drive was busy remapping, the symptoms would match.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
On 03/16/2011 05:38 PM, Carlos E. R. wrote:
> On 2011-03-16 23:06, quackers wrote:
>> And if it just continues without change? Just wait for the hard drive
>> to die?
>
> Are you suspecting the drive is bad?
>
> If your drive was busy remapping, the symptoms would match.
>
As I will probably appear to be condescending, I’ll only post once. Switch to
Virtual Terminal 10 to see what errors are there. As you are an expert, I know
that you will already know how to do that.
Iwfinger,
I have no reason to suspect that a hard drive is bad. The smart status is healthy on both drives.
You didn’t appear condescending at all, until the last paragraph.
Perhaps you guys may want to cool down a bit. DenverD is being his usual self, if you don’t like just ignore him, he will go away eventually - although he is helpful most of the time, even if in a hurtful sort of way
Back to your problem I’d speculate the repartitioning triggered a filesystem check that should have been completed before loading level 5, but for some reason it kept going on - perhaps something in the new run-parallel-to-speed-boot changes in 11.4. If it kept on, you killing the machine would just trigger it back in the subsequent runs.
Just to be sure I’d try a HD check with the pmagic liveCD, it has many test tools including smart HD tests you can run.
Thank you for your comments, brunomcl.
I’m sure DenverD is normally helpful. I’m even prepared to consider that he thought that he was being helpful to me. It just didn’t read that well, to me.
The matter is forgotten at this end.
Thank you also for reading my first post fully before commenting.
With regard to the file system check, that was indeed run during the text stage of booting. The test was completed without reporting any problems.
My main concern is that something appeared to be running which was not picked up by System Monitor. This has not happened before on any Linux system I have run. It is something I expect of a Windows system, but not a Linux system.
As a precautionary measure I wil check the status of my HDD’s.
Thanks for your time.
I also suspect a fsck process - unfortunately the SuSE-fsck is kinda quite, even when no bootsplash is running. I suggest to manually repeat that fsck in a more verbose way like this:
fsck -Cv /dev/sdxn
…where ‘/dev/sdxn’ is the respective partition you would like to check (/dev/sda1 for example). You should run this on both the /- and the /home-partition (and any other partition present). This should be done via a live-system such as Knoppix.
Note that fsck is checking for problems within the filesystem, not the disk itself (as s.m.a.r.t. does). A disk can be healthy while the filesystem is mashed.
Thanks for the suggestion gropiuskalle. I ran the file check on sdb8 (/) and sdb9 (/home). They both completed in seconds and reported them clean.
I booted back in to 11.4 and the same problem occurred. After opening System Monitor, which reported no anomalies in either cpu usage or ram usage, I then ran “top” from the terminal. Rather than waiting for it to open I went outside for a smoke
When I returned, top reported that the cpu was barely being used, but the ram (2GB) was completely used up by the “prepare_preload” process.
I have googled this process and found one or two bugzilla reports about it causing problems in the past. One way to stop this happening was to remove the “preload” package and lock it, to prevent it from re-installing itself. That seems a little drastic, but I’m prepared to do it if I am advised in that way. I believe it just assists boot times by preloading stuff - not a must-have then really, for me.
Has anyone else had this problem and fixed it?
It would also be nice to have some views as to why System Monitor reported that only approximately 290MB of ram was being used - when quite clearly it was maxed out!
Thanks.
On 2011-03-17 07:36, quackers wrote:
> When I returned, top reported that the cpu was barely being used, but
> the ram (2GB) was completely used up by the “prepare_preload” process.
Interesting.
> I have googled this process and found one or two bugzilla reports about
> it causing problems in the past.
I think you should add a new bugzilla.
> It would also be nice to have some views as to why System Monitor
> reported that only approximately 290MB of ram was being used - when
> quite clearly it was maxed out!
No idea, I always use ‘top’ myself.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
On Thu, 17 Mar 2011 15:50:06 +0530, Carlos E. R.
<robin_listas@no-mx.forums.opensuse.org> wrote:
>> It would also be nice to have some views as to why System Monitor
>> reported that only approximately 290MB of ram was being used - when
>> quite clearly it was maxed out!
> No idea, I always use ‘top’ myself.
i often find discrepancies between ksysguard & top. interestingly, the
“system load viewer” plasmoid (graphical representation of CPU(s), memory,
and swap) seems to show the real load rather than what ksysguard reports.
so i keep that open in a panel, and when the colored graphs show more than
usual activity i investigate with top or htop.
–
phani.
On 03/17/2011 01:36 AM, quackers wrote:
>
> Thanks for the suggestion gropiuskalle. I ran the file check on sdb8 (/)
> and sdb9 (/home). They both completed in seconds and reported them
> clean.
> I booted back in to 11.4 and the same problem occurred. After opening
> System Monitor, which reported no anomalies in either cpu usage or ram
> usage, I then ran “top” from the terminal. Rather than waiting for it to
> open I went outside for a smoke
> When I returned, top reported that the cpu was barely being used, but
> the ram (2GB) was completely used up by the “prepare_preload” process.
> I have googled this process and found one or two bugzilla reports about
> it causing problems in the past. One way to stop this happening was to
> remove the “preload” package and lock it, to prevent it from
> re-installing itself. That seems a little drastic, but I’m prepared to
> do it if I am advised in that way. I believe it just assists boot times
> by preloading stuff - not a must-have then really, for me.
> Has anyone else had this problem and fixed it?
> It would also be nice to have some views as to why System Monitor
> reported that only approximately 290MB of ram was being used - when
> quite clearly it was maxed out!
For the record, preload is not needed. I do driver development and 99.99% of the
time, I am not running the standard kernel. Non of the ones that I build have
that “feature”, and I assure you that my systems work fine.
What does “free” have to say about your memory?
By all means, file a bug report if there is not already one.
Thanks for the extra info, people
lwfinger, as you say, the previous bugs were quite old. One of the “fixes” in one of them was to remove the preload package. I have no qualms about doing that, as I’m confident that it won’t be badly missed It does seem to be a little drastic, but, no problem.
I have reported a bug. I think I’ve done it correctly, but as it’s my first bug with opensuse, I’m not absolutely certain.
https://bugzilla.novell.com/show_bug.cgi?id=680456
While the problem exists top reports no free ram. System Monitor reported just 290MB of 2GB being used.
On 03/17/2011 03:36 PM, quackers wrote:
>
> While the problem exists top reports no free ram. System Monitor
> reported just 290MB of 2GB being used.
that could/should be your second bug report…
–
DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.1.8, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11
On 2011-03-17 15:36, quackers wrote:
> While the problem exists top reports no free ram. System Monitor
> reported just 290MB of 2GB being used.
Are you sure that missing memory is not in buffers? It may be that SM
doesn’t report memory in buffers or cache.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)