[opensuse] My desktop has become "lazy" lately
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, That's the best description, its become lazy. Some times, if I'm not typing or moving the mouse, the entire machine stops. I see the display of gkrelmn stop. I was calculating the size of a directory using 'mc', went out for an hour, and the thing was exactly as I had left it: no work done at all. I move the mouse, and it suddenly starts silently working again for a few seconds, then stop. I go to ctrl-alt-f1, start there "top", and it works. I go back to the desktop (gnome) and the laziness seems gone. As I'm writing this, I stop that 'top', and the desktop continues working - no, it doesn't, it stops after 10" or so. I set the clock to show seconds, it works. No seconds, it stops - but not always. What is happening? Kernel? Desktop? Goblins? - -- Cheers, Carlos Robinson -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZRQRtTMYHG2NR9URAhEaAJ4+kZrWoHdD3RRehQ5N7w5yBVoAqQCglje3 QPHIdDKXSx0TJl+VLeca6HI= =adum -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Sounds to me like your system is only working when being external interrupts are present. How to solve, beats me. -----Original Message----- From: Carlos E. R. [mailto:robin.listas@telefonica.net] Sent: Monday, 17 December 2007 1:03 a.m. To: OS Subject: [opensuse] My desktop has become "lazy" lately -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, That's the best description, its become lazy. Some times, if I'm not typing or moving the mouse, the entire machine stops. I see the display of gkrelmn stop. I was calculating the size of a directory using 'mc', went out for an hour, and the thing was exactly as I had left it: no work done at all. I move the mouse, and it suddenly starts silently working again for a few seconds, then stop. I go to ctrl-alt-f1, start there "top", and it works. I go back to the desktop (gnome) and the laziness seems gone. As I'm writing this, I stop that 'top', and the desktop continues working - no, it doesn't, it stops after 10" or so. I set the clock to show seconds, it works. No seconds, it stops - but not always. What is happening? Kernel? Desktop? Goblins? - -- Cheers, Carlos Robinson -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZRQRtTMYHG2NR9URAhEaAJ4+kZrWoHdD3RRehQ5N7w5yBVoAqQCglje3 QPHIdDKXSx0TJl+VLeca6HI= =adum -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sunday 16 December 2007 20:18, Philip Dowie wrote:
Sounds to me like your system is only working when being external interrupts are present. How to solve, beats me.
-----Original Message----- From: Carlos E. R. [mailto:robin.listas@telefonica.net] Sent: Monday, 17 December 2007 1:03 a.m. To: OS Subject: [opensuse] My desktop has become "lazy" lately
Hi,
That's the best description, its become lazy. Some times, if I'm not typing or moving the mouse, the entire machine stops. I see the display of gkrelmn stop. I was calculating the size of a directory using 'mc', went out for an hour, and the thing was exactly as I had left it: no work done at all.
I move the mouse, and it suddenly starts silently working again for a few seconds, then stop.
I go to ctrl-alt-f1, start there "top", and it works. I go back to the desktop (gnome) and the laziness seems gone. As I'm writing this, I stop that 'top', and the desktop continues working - no, it doesn't, it stops after 10" or so.
I set the clock to show seconds, it works. No seconds, it stops - but not always.
What is happening? Kernel? Desktop? Goblins?
-- Cheers, Carlos Robinson
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Revert back to runlevel 3 and see if the system still is "lazy". Check the syslog for errors. Report back. -- /Rikard ----------------------------------------------------------------------------- email : rikard.j@rikjoh.com web : http://www.rikjoh.com mob: : +46 (0)763 19 76 25 ------------------------ Public PGP fingerprint ---------------------------- < 15 28 DF 78 67 98 B2 16 1F D3 FD C5 59 D4 B6 78 46 1C EE 56 >
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2007-12-16 at 20:29 +0100, Rikard Johnels wrote:
Revert back to runlevel 3 and see if the system still is "lazy". Check the syslog for errors.
Report back.
If I go to a text console, in level 5, it appears not to go lazy, but it is difficult to say because the laziness is always erratic, doesn't happen always. But so far I have never seen it in text mode. Right now, I can't trigger the lazy behavior in any mode. When I wrote the message I got the laziness response very easily, but there was never any message in the logs, nothing at all. I'll have to wait till another day when this happens again. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZbwktTMYHG2NR9URAhrbAKCNu3X0e/H64FCvPj6is5V90wuyKACeJ9BD q/mSDWgT5v+JNxmoWrBtdyA= =gCKR -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
I go to ctrl-alt-f1, start there "top", and it works. I go back to the desktop (gnome) and the laziness seems gone. As I'm writing this, I stop that 'top', and the desktop continues working - no, it doesn't, it stops after 10" or so.
I set the clock to show seconds, it works. No seconds, it stops - but not always.
What is happening? Kernel? Desktop? Goblins?
It _sounds_ like some power-saving feature is set to become "active" a short time after last user input (also sounds like it might be messing up; but it could be a 'feature' for some people (on laptop, on battery power - go to 'standby' 10 seconds after last mouse or keyboard input? Dunno... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2007-12-16 at 14:28 -0800, Linda Walsh wrote:
It _sounds_ like some power-saving feature is set to become "active" a short time after last user input (also sounds like it might be messing up; but it could be a 'feature' for some people (on laptop, on battery power - go to 'standby' 10 seconds after last mouse or keyboard input? Dunno...
Good idea... but it doesn't appear to be that easy. In the "Power management preferences" window, the action "put computer to sleep when inactive for:" is set to "Never" (and the minimum period possible is 11 minutes). And both the power and suspend buttons are set to trigger "hibernate". - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZbm1tTMYHG2NR9URAmh4AJ0S/potXl3MpF3AYqaqg8gq7XA0uwCbBa2z 8cKu4qcHeLjvdj+uth89jNo= =Tblm -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
The Sunday 2007-12-16 at 14:28 -0800, Linda Walsh wrote:
It _sounds_ like some power-saving feature is set to become "active" a short time after last user input (also sounds like it might be messing up; but it could be a 'feature' for some people (on laptop, on battery power - go to 'standby' 10 seconds after last mouse or keyboard input? Dunno...
Good idea... but it doesn't appear to be that easy.
Never is :-(. It would likely be at the kernel level - like the kernel is going into a "low-load" setting -- since your system isn't suspending or going to sleep. There are some cpu scheduling modules in the kernel (util 'powertop' makes suggestions for modules to include for laptop or low power systems. Something the "cpufreq_ondemand" module in the in a suse stock kernel (2.6.18.2-34) I see cpufreq modules for conservative, ondemand, powersave. In some kernel version (gee, am just so specific -- I read information in greater amounts than my brain auto-indexes... (:-)) I thought there was a bug in some later kernel version concerning the new tickless kernel and the ondemand-cpu module, but I don't think the tickless patches are in 2.6.18. I wonder if there are any ways to tweak those modules -- but it appears they would only be used on a system that has a variable cpu frequency -- so unless you have that hardware....(what hw did you say you had?)... Good luck... L -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2007-12-16 at 16:28 -0800, Linda Walsh wrote:
Good idea... but it doesn't appear to be that easy.
Never is :-(.
It would likely be at the kernel level - like the kernel is going into a "low-load" setting -- since your system isn't suspending or going to sleep. There are some cpu scheduling modules in the kernel (util 'powertop' makes suggestions for modules to include for laptop or low power systems. Something the "cpufreq_ondemand" module in the in a suse stock kernel (2.6.18.2-34) I see cpufreq modules for conservative, ondemand, powersave.
In some kernel version (gee, am just so specific -- I read information in greater amounts than my brain auto-indexes... (:-))
Doesn't happen to us all? :-)
I thought there was a bug in some later kernel version concerning the new tickless kernel and the ondemand-cpu module, but I don't think the tickless patches are in 2.6.18.
tickless... sounds familiar.
I wonder if there are any ways to tweak those modules -- but it appears they would only be used on a system that has a variable cpu frequency -- so unless you have that hardware....(what hw did you say you had?)...
I'm not sure if my hardware has that capacity. It is a pentium IV @ 1800, single cpu. Lets see what /proc says: nimrodel:~/notas # cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 1 model name : Intel(R) Pentium(R) 4 CPU 1.80GHz stepping : 2 cpu MHz : 1800.190 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm bogomips : 3603.26 clflush size : 64 Then, in /proc/acpi/processor/CPU0/* info: processor id: 0 acpi id: 1 bus mastering control: no power management: yes throttling control: yes limit interface: yes limit: active limit: P0:T0 user limit: P0:T0 thermal limit: P0:T0 power: active state: C2 max_cstate: C8 bus master activity: fb7ef96b maximum allowed latency: 6666 usec states: C1: type[C1] promotion[C2] demotion[--] latency[000] usage[16988520] duration[00000000000000000000] *C2: type[C2] promotion[--] demotion[C1] latency[090] usage[42284950] duration[00000000738930421870] throttling: state count: 2 active state: T0 states: *T0: 00% T1: 50% My suspect is the clock. I was having problems with the default system clock, which is 'acpi_pm', which had delays of several minutes per hour. I changed to 'tsc', which works, although the kernel complains during boot: nimrodel:~ # grep -i "clock\|tsc\|TSC" /var/log/boot.msg <6>Time: tsc clocksource has been installed. <6>Real Time Clock Driver v1.12ac <4>Marking TSC unstable due to: possible TSC halt in C2. <=== <6>Time: acpi_pm clocksource has been installed. <6>intel8x0_measure_ac97_clock: measured 50881 usecs <6>intel8x0: clocking to 48000 doneSetting up the hardware clockdone And: Dec 8 13:56:52 nimrodel kernel: Clocksource tsc unstable (delta = 32800377181 ns) However, I think I had instances of these freezing episodes weeks ago, before I set the 'tsc' clock, so... I have no idea. Another symptom: Twice, completely randomly, the system froze. The last time was several days ago. The keyboard stopped responding, the mouse, the display... I thought of powering another machine and entering through ssh; but the moment the other machine had finished booting up, my main machine continued working as if nothing had happened. This has happened twice, and it is very weird. The only thing I know for sure is that it is a software issue of suse 10.3: they started the day after I upgraded. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZc4vtTMYHG2NR9URAonwAJ4gAlNIHos4nyPHbHoIshnzmAC9swCgiFMH dq+aqxLD/qrkqBmEhyAJjuE= =SkAc -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Mon, 2007-12-17 at 02:17 +0100, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Sunday 2007-12-16 at 16:28 -0800, Linda Walsh wrote:
However, I think I had instances of these freezing episodes weeks ago, before I set the 'tsc' clock, so...
I have no idea.
Another symptom:
Twice, completely randomly, the system froze. The last time was several days ago. The keyboard stopped responding, the mouse, the display... I thought of powering another machine and entering through ssh; but the moment the other machine had finished booting up, my main machine continued working as if nothing had happened.
This has happened twice, and it is very weird. The only thing I know for sure is that it is a software issue of suse 10.3: they started the day after I upgraded.
This thread is sounding suspiciously similar to my drive failure Thursday, following a YOU last weekend. See my thread "ata2 suddenly bad". I mean the randomly "sluggish" aka "lock-up" syndrome. I have since installed an EIDE drive with fresh install of "boxed 10.3", reloaded all my original apps, and have done a new YOU (with the same kernel version as last week). So far...28 hrs and counting...all is fabulous. I can't add anything else, except that there have been several weird threads this week about similar "stall" symptoms...one re oocalc, mine, yours, and a couple others. I had thought mine was due to Beagle, or perhaps because it was an "upgrade" from 10.2. Then I lost both /sda8 and /sdb2...something is amiss somewhere. Or not...I might have had a failing SATA drive...I am currently not using any of /sdb partitions. Tom in NM -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2007-12-16 at 21:11 -0700, Tom Patton wrote:
This thread is sounding suspiciously similar to my drive failure Thursday, following a YOU last weekend. See my thread "ata2 suddenly bad". I mean the randomly "sluggish" aka "lock-up" syndrome.
Yes, I read it. I couldn't decide if it is significantly similar or chance.
I have since installed an EIDE drive with fresh install of "boxed 10.3", reloaded all my original apps, and have done a new YOU (with the same kernel version as last week). So far...28 hrs and counting...all is fabulous.
I can't add anything else, except that there have been several weird threads this week about similar "stall" symptoms...one re oocalc, mine, yours, and a couple others.
Yes.
I had thought mine was due to Beagle, or perhaps because it was an "upgrade" from 10.2. Then I lost both /sda8 and /sdb2...something is amiss somewhere. Or not...I might have had a failing SATA drive...I am currently not using any of /sdb partitions.
Dificult to say, yes. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZoV+tTMYHG2NR9URAszFAJ46lzJZcBzq64wGhb772h5ovmJ6aQCdFhr+ CwzO90fnazS/KGwClW8VQjk= =+q0Z -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Mon, 2007-12-17 at 15:19 +0100, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Dificult to say, yes.
- -- Cheers, Carlos E. R.
I know...coincidence or the twilight zone...after 52 hours, all is well here, on one EIDE and one SADA drive. Only thing NOT the same, is that I did NOT install either KDE or GNOME desktop environs this time...and no BEAGLE. Other than that, all upgrades are at the same level, and all apps are installed, but using strictly FVWM. If it continues a couple weeks, I might install KDE again. But by then, my wife probably won't miss it...;-) Good luck on your gremlin, Carlos! Tom in NM -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
I have seen this type of behavior once before on one of our Sparc boxes. It would randomly stop processing, then wake up. You could watch the clock stop ticking and then start again. Some applications (that weren't looking for a clock tick) would be okay, but others (like Kerberos, for some reason) would stop responding. In that case it ended up being the onboard RTC. We had Sun swap out the system board and the box has been stable ever since. -jc -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2007-12-17 at 06:04 -0700, Jc Polanycia wrote:
I have seen this type of behavior once before on one of our Sparc boxes. It would randomly stop processing, then wake up. You could watch the clock stop ticking and then start again. Some applications (that weren't looking for a clock tick) would be okay, but others (like Kerberos, for some reason) would stop responding. In that case it ended up being the onboard RTC. We had Sun swap out the system board and the box has been stable ever since.
But I refuse to believe it can be a hardware problem, because 10.2 had no problems, and the problems started as soon as I upgraded to 10.3 - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZoXGtTMYHG2NR9URAqm5AJ4tsFGFDrk5nLXoz0+3/8i04mB7EACglI3z U6j0YK6bSL+JHmZkbmc8hVQ= =kyF9 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 12/17/2007 10:20 PM, Carlos E. R. wrote:
The Monday 2007-12-17 at 06:04 -0700, Jc Polanycia wrote:
I have seen this type of behavior once before on one of our Sparc boxes. It would randomly stop processing, then wake up. You could watch the clock stop ticking and then start again. Some applications (that weren't looking for a clock tick) would be okay, but others (like Kerberos, for some reason) would stop responding. In that case it ended up being the onboard RTC. We had Sun swap out the system board and the box has been stable ever since.
But I refuse to believe it can be a hardware problem, because 10.2 had no problems, and the problems started as soon as I upgraded to 10.3
Maybe it is not indicating a hardware failure, but the source of the problem. It sounds like it is related to your clock problem. Perhaps tsc is often marked as unreliable for a reason. Maybe try jiffies, since the acpi_pm seemed to cause your clock even more problems. -- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-12-18 at 07:27 +0800, Joe Morris (NTM) wrote:
But I refuse to believe it can be a hardware problem, because 10.2 had no problems, and the problems started as soon as I upgraded to 10.3
Maybe it is not indicating a hardware failure, but the source of the problem. It sounds like it is related to your clock problem. Perhaps tsc is often marked as unreliable for a reason. Maybe try jiffies, since the acpi_pm seemed to cause your clock even more problems.
The "acpi_pm" is what the kernel chooses, but doesn't work well. Yep, I could try "jiffies", I guess... It will take days to make sure if it helps. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZyL0tTMYHG2NR9URAnF9AJ9D5vwseoeKXDZQrKB4gWnU7o4X8QCfYT+9 1is50ZEqq2+udBeeCwoGenA= =VxbD -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-12-18 at 07:27 +0800, Joe Morris (NTM) wrote:
But I refuse to believe it can be a hardware problem, because 10.2 had no problems, and the problems started as soon as I upgraded to 10.3
Maybe it is not indicating a hardware failure, but the source of the problem. It sounds like it is related to your clock problem. Perhaps tsc is often marked as unreliable for a reason. Maybe try jiffies, since the acpi_pm seemed to cause your clock even more problems.
I just tried, and my system crashed: echo jiffies > /sys/devices/system/clocksource/clocksource0/current_clocksource I tried again, in runlevel 3, and I discovered that the clock stopped. I tried again to go back to tsc, but the clock continued stopped. I tried to reboot, and reboot crashed. Is no good. * acpi_pm looses time, like several minutes per hour. * tsc "appears" to work well, but might be suspicious * jiffies doesn't even work, time stops, applications depending on the clock stop - even "halt" stops when issuing a beep because the beep never times out. * pit I haven't tried. Should I? :-/ - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHaVXPtTMYHG2NR9URAvzNAJ9acZlnuG3Q01zbzJbc8HcXsGW15ACgiiYt wsXucVfkphxbySqm4FgaQFY= =VMjg -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 12/20/2007 01:32 AM, Carlos E. R. wrote:
I just tried, and my system crashed:
echo jiffies > /sys/devices/system/clocksource/clocksource0/current_clocksource
I tried again, in runlevel 3, and I discovered that the clock stopped. I tried again to go back to tsc, but the clock continued stopped. I tried to reboot, and reboot crashed.
Is no good.
* acpi_pm looses time, like several minutes per hour.
* tsc "appears" to work well, but might be suspicious According to an error message on my system, I think your clock and "lazy desktop" are the same core problem. From your other posts, it appeared 10.3 introduced processor frequency control to your system. From my system, it says something to the effect, marking tsc as unstable due to cpufreq changes. My CPU does have Cool and Quiet, so it sounds correct. Mine uses therefore acpi_pm. jmorris:/home/joe # cat /sys/devices/system/clocksource/clocksource0/current_clocksource acpi_pm I would therefore suggest to pass cpufreq=no as a boot argument, and use tsc, to see if that allows your clock to work correctly, without a cpu frequency change perhaps causing a "lazy desktop" spell.
* jiffies doesn't even work, time stops, applications depending on the clock stop - even "halt" stops when issuing a beep because the beep never times out. Ouch. Sorry. I based that on what a few others wrote in. I had no idea it would cause such problems on your system.
* pit I haven't tried. Should I?
After what jiffies did, I wouldn't. Is this your system with xfs? Just trying not to offer bad advice. -- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-12-20 at 08:09 +0800, Joe Morris (NTM) wrote:
Is no good.
* acpi_pm looses time, like several minutes per hour.
* tsc "appears" to work well, but might be suspicious According to an error message on my system, I think your clock and "lazy desktop" are the same core problem. From your other posts, it appeared 10.3 introduced processor frequency control to your system. From my system, it says something to the effect, marking tsc as unstable due to cpufreq changes. My CPU does have Cool and Quiet, so it sounds correct. Mine uses therefore acpi_pm. jmorris:/home/joe # cat /sys/devices/system/clocksource/clocksource0/current_clocksource acpi_pm I would therefore suggest to pass cpufreq=no as a boot argument, and use tsc, to see if that allows your clock to work correctly, without a cpu frequency change perhaps causing a "lazy desktop" spell.
Could be...
* jiffies doesn't even work, time stops, applications depending on the clock stop - even "halt" stops when issuing a beep because the beep never times out. Ouch. Sorry. I based that on what a few others wrote in. I had no idea it would cause such problems on your system.
No, no, don't worry, I wanted to do that test before you mentioned it. I have had worse crashes than that one, I have another problem with encrypted filesystem causing a crash. I may report the problem with jiffies as a bug to bugzilla.
* pit I haven't tried. Should I?
After what jiffies did, I wouldn't. Is this your system with xfs? Just trying not to offer bad advice.
As a matter of fact, I do have some xfs partitions in my system, but not the root. Why? Do you think that clock problems can cause problems in xfs? That does interests me a lot, because the encrypted filesystem I mentioned above is xfs. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHabZ6tTMYHG2NR9URAgOyAJsFBw8qcyATfr+bSIiwgeG4FWvZrgCfVKnU 36U6LSN7McpDtpeSMmCj8iM= =YpBY -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 12/20/2007 08:25 AM, Carlos E. R. wrote:
After what jiffies did, I wouldn't. Is this your system with xfs? Just trying not to offer bad advice.
As a matter of fact, I do have some xfs partitions in my system, but not the root. Why? Do you think that clock problems can cause problems in xfs? That does interests me a lot, because the encrypted filesystem I mentioned above is xfs.
No, nothing like that. I was just thinking of the effect of those kinds of crashes on xfs. I would feel really bad if the advice I gave caused a crash that corrupted a filesystem. -- Joe Morris Registered Linux user 231871 running openSUSE 10.3 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-12-20 at 08:53 +0800, Joe Morris (NTM) wrote:
On 12/20/2007 08:25 AM, Carlos E. R. wrote:
After what jiffies did, I wouldn't. Is this your system with xfs? Just trying not to offer bad advice.
As a matter of fact, I do have some xfs partitions in my system, but not the root. Why? Do you think that clock problems can cause problems in xfs? That does interests me a lot, because the encrypted filesystem I mentioned above is xfs.
No, nothing like that. I was just thinking of the effect of those kinds of crashes on xfs. I would feel really bad if the advice I gave caused a crash that corrupted a filesystem.
Ah... I see. Pity. :-} I was hopping we had unearthed a glimpse of hope to my other big problem. My XFS partitions are proving quite resilient to those crashes, so far. I'll keep my finger crossed. The two crashes caused be the clock change were pretty mild, actually. The other crashes, triggered by large file writing to an encrypted filesystem, are much worse, because it directly affects disk writing, and I loose things like bash history (nothing serious, just inconvenient). However, the encrypted partition itself seems unaffected. Very weird, and no clues. No log entries. The XFS filesystem uses some realtime routines, there is a mention of this in the manuals somewhere. I was hopping that the clock being bad could have some known effect on it. Who knows... I'll have to try cpufreq=no soon. I'll write it up in menu.lst before I forget [...] done. (I did a kernel recompile this afternoon and I want to see if it has any effect, first) - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHadK+tTMYHG2NR9URAjUGAJ98AwhMwUwfQMgVLCsc9DMvZUWpiwCfW/YI 8BNuO1pY6il/wf82nyo5ECk= =FTmZ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-12-20 at 08:09 +0800, Joe Morris (NTM) wrote:
I would therefore suggest to pass cpufreq=no as a boot argument, and use tsc, to see if that allows your clock to work correctly, without a cpu frequency change perhaps causing a "lazy desktop" spell.
I forgot a detail. I wonder if it is possible to do that while running, by echoing a value to proc or sys somewhere. The throttling state can be read from /proc/acpi/processor/CPU0/throttling, but I don't think that's the appropiate place to write to, looking at the format: state count: 2 active state: T0 states: *T0: 00% T1: 50% - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHabo+tTMYHG2NR9URAsHvAJ9ALrMSwz7XauV2FhRiNwK2ciuZwQCfebZ5 QoiPF5I6tasZL9N+klX1Iso= =qUtv -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Thursday 2007-12-20 at 08:09 +0800, Joe Morris (NTM) wrote:
I would therefore suggest to pass cpufreq=no as a boot argument, and use
Tried that, yesterday. All went fine, till a minute ago, it freezed. A key (shift), and it waked up. So, it is not that... And no mistake, frequency has been disssabled: Dec 21 21:01:11 nimrodel powersaved[4736]: WARNING (CpufreqManagement:51) No capability cpufreq_control Dec 21 21:01:11 nimrodel powersaved[4736]: WARNING (CpufreqManagement:51) No capability cpufreq_control I think I'll try "NO_HZ", somebody mentioned it is problematic. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHbWhrtTMYHG2NR9URAi7fAJ9li3ogxoGLpi8Qc0NpJ1MKyTE7UQCfTDTj cO1w3CN/9tXlM6Djar2JEPY= =fBvQ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Saturday 2007-12-22 at 20:41 +0100, I wrote:
The Thursday 2007-12-20 at 08:09 +0800, Joe Morris (NTM) wrote:
I would therefore suggest to pass cpufreq=no as a boot argument, and use
Tried that, yesterday. All went fine, till a minute ago, it freezed. A key (shift), and it waked up. So, it is not that...
And no mistake, frequency has been disssabled:
Dec 21 21:01:11 nimrodel powersaved[4736]: WARNING (CpufreqManagement:51) No capability cpufreq_control Dec 21 21:01:11 nimrodel powersaved[4736]: WARNING (CpufreqManagement:51) No capability cpufreq_control
I think I'll try "NO_HZ", somebody mentioned it is problematic.
Well, I recompiled the kernel with that option dissabled: no difference. One thing I noticed: a ping from another computer in the LAN awakes it, and keeps it from dozing. I stops the pings, and a few seconds later, my PC dozes away. It needs interrupts comming in... - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHcV88tTMYHG2NR9URAiSIAJwM+axs9JEQpJW8uFQDrJHcwL9mOQCeIp+e NcishfuYFv9GRGCcTWRPe5s= =CQEW -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
The Sunday 2007-12-16 at 14:28 -0800, Linda Walsh wrote:
It _sounds_ like some power-saving feature is set to become "active" a short time after last user input (also sounds like it might be messing up; but it could be a 'feature' for some people (on laptop, on battery power - go to 'standby' 10 seconds after last mouse or keyboard input? Dunno...
Good idea... but it doesn't appear to be that easy. In the "Power management preferences" window, the action "put computer to sleep when inactive for:" is set to "Never" (and the minimum period possible is 11 minutes). And both the power and suspend buttons are set to trigger "hibernate".
-- Cheers, Carlos E. R.
Try setting "Never" to something else, save, exit, then change it back to "Never" and see if that help. Other than that, just set the clock to show seconds and forget about it ;-) -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Sunday 2007-12-16 at 21:29 -0600, David C. Rankin wrote:
inactive for:" is set to "Never" (and the minimum period possible is 11 minutes). And both the power and suspend buttons are set to trigger "hibernate".
Try setting "Never" to something else, save, exit, then change it back to "Never" and see if that help.
I already did that. Difficult to say yet if it helps or not - the problem is very random.
Other than that, just set the clock to show seconds and forget about it ;-)
Tried that too. But gkrelmn does updates to its graphs every second. I notice the problem by watching these graphs ans eeing them freeze. Ie, having a program update the display every second does not help. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZoUatTMYHG2NR9URArNHAJ0WfBT8BwiQ+52pF3muwFN0Au5G1ACffcaF w0FGPYq1xgKQheLuEiXSRFg= =jame -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Monday 2007-12-17 at 15:18 +0100, Carlos E. R. wrote:
The Sunday 2007-12-16 at 21:29 -0600, David C. Rankin wrote:
Other than that, just set the clock to show seconds and forget about it ;-)
Tried that too.
But gkrelmn does updates to its graphs every second. I notice the problem by watching these graphs ans eeing them freeze. Ie, having a program update the display every second does not help.
I have the clock showing seconds since yesterday. I was writing emails, pressed enter and waited till the response of the machine... and it stopped, for something like 8 seconds. It came alive again without me touching anything. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZ5/PtTMYHG2NR9URAiTkAJ40reJyUXaZwLIVtefKYqlI7mbI8gCeIA2x XtlUwuSCCzUHc2NKeeuHi84= =/zP1 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
The Monday 2007-12-17 at 15:18 +0100, Carlos E. R. wrote:
The Sunday 2007-12-16 at 21:29 -0600, David C. Rankin wrote:
Other than that, just set the clock to show seconds and forget about it ;-)
Tried that too.
But gkrelmn does updates to its graphs every second. I notice the problem by watching these graphs ans eeing them freeze. Ie, having a program update the display every second does not help.
I have the clock showing seconds since yesterday.
I was writing emails, pressed enter and waited till the response of the machine... and it stopped, for something like 8 seconds. It came alive again without me touching anything.
-- Cheers, Carlos E. R.
Carlos, There has to be an interrupt sharing conflict somewhere. Can you disable some hardware and see if it stops. (unplug usb devices, pull the sound card, etc..) If it does, then add the pieces back one at a time. It might even be a pci bus conflict where moving cards to a different might solve it. (note theses are just educated guesses) -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-12-18 at 07:36 -0600, David C. Rankin wrote:
Carlos,
There has to be an interrupt sharing conflict somewhere. Can you disable some hardware and see if it stops. (unplug usb devices, pull the sound card, etc..) If it does, then add the pieces back one at a time. It might even be a pci bus conflict where moving cards to a different might solve it. (note theses are just educated guesses)
The only hardware change has been removal of a TV card, the rest is the same in months, no, years. And the removal of that card didn't alter this behaviour. No, this is a software issue, because they started the same day I installed 10.3. I wrote a script: #!/bin/bash while true ; do date +"%T" >> /home/cer/marca.log sleep 1 done and had it running from a text terminal in alt-ctrl-f2. I looked just after one of those small freezes, and the log showed this: 12:41:21 12:41:22 12:41:23 12:41:24 12:41:47 <=== 12:41:48 12:41:49 12:41:50 A time jump! The machine stoped working, but the time was correctly updated when activity returned. And, I know it is not the desktop, because this script was not running under the desktop. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHZ+eUtTMYHG2NR9URAqyVAKCPeLeBmoZ1z12H+q3nW9kPs9jDigCeOqbL ks2LjfzsGmebGcIXoV2mScI= =IN8N -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tuesday 18 December 2007 16:30, Carlos E. R. wrote:
The Tuesday 2007-12-18 at 07:36 -0600, David C. Rankin wrote:
Carlos,
There has to be an interrupt sharing conflict somewhere. Can you disable some hardware and see if it stops. (unplug usb devices, pull the sound card, etc..) If it does, then add the pieces back one at a time. It might even be a pci bus conflict where moving cards to a different might solve it. (note theses are just educated guesses)
The only hardware change has been removal of a TV card, the rest is the same in months, no, years. And the removal of that card didn't alter this behaviour.
No, this is a software issue, because they started the same day I installed 10.3.
I wrote a script:
#!/bin/bash while true ; do date +"%T" >> /home/cer/marca.log sleep 1 done
and had it running from a text terminal in alt-ctrl-f2. I looked just after one of those small freezes, and the log showed this:
12:41:21 12:41:22 12:41:23 12:41:24 12:41:47 <=== 12:41:48 12:41:49 12:41:50
A time jump! The machine stoped working, but the time was correctly updated when activity returned.
And, I know it is not the desktop, because this script was not running under the desktop.
-- Cheers, Carlos E. R.
What services etc do you have running? What software is running? Is the machine doing heavy disk accessing? Maybe swapping memory to disk? -- /Rikard ----------------------------------------------------------------------------- email : rikard.j@rikjoh.com web : http://www.rikjoh.com mob: : +46 (0)763 19 76 25 ------------------------ Public PGP fingerprint ---------------------------- < 15 28 DF 78 67 98 B2 16 1F D3 FD C5 59 D4 B6 78 46 1C EE 56 >
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Tuesday 2007-12-18 at 17:53 +0100, Rikard Johnels wrote:
What services etc do you have running? What software is running?
The default services that come with the distro, mostly. Nothing weird. This is a desktop machine. There is amavis, postfix, spamd...
Is the machine doing heavy disk accessing? Maybe swapping memory to disk.
No, no. And certainly not when these "lapsus" occur. Swap is used for hibernation. But I'll watch the swap graph in gkrellm the next time one of those lapsus occur. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHaCE6tTMYHG2NR9URAiHcAJ4t+FouTNZhcN6hcnx6EJV2F8L20wCfcSuc vVp9OZPQKaKyx9X9BdtT+30= =k0Rp -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (8)
-
Carlos E. R. -
David C. Rankin -
Jc Polanycia -
Joe Morris (NTM) -
Linda Walsh -
Philip Dowie -
Rikard Johnels -
Tom Patton