You are not logged in.
Have you added the "noatime" option to your fstab ?
Offline
Yes noatime also was added.
Offline
Just did a bit more research. I'm gonna keep track track of load cycles for the next couple of days in xp and arch and will post how it goes.
Offline
It looks like the load unload cycles is a real problem for me, I installed smartmontools in xp and arch after reading a different site.
Before I shutdown xp, I had 677 Load_Cycle_Count on my drive.
I booted into Arch and just did some browsing, only running X, fluxbox and opera.
In the next 24 minutes the HDD did 41 cycles, which is way way too many (it should be less than 15/hour, 15/hour equates to about 4 years battery life from what I've read), the noise I mentioned earlier is actually the noise of load/unload as after every time I heard it the count went +1.
I've since booted back into XP ( and have had only one load cycles in 20 minutes)
Can a few people post their load cycles stats. You need to install smartmontools if you haven't got it already and run (where /dev/hda is the disk you are checking)
smartctl -a /dev/hda | grep Load_Cycle_Count
This shows the load cycle count.
Also if you run
smartctl -a /dev/hda | grep Power_On_Hours
gives you the amount of hours your laptop has been on so you can work out how many loads you are getting per hour.
Last edited by snappy (2008-12-19 05:52:21)
Offline
Have you looked at this thread: http://bbs.archlinux.org/viewtopic.php?id=39258?
It seems to be a problem without a simple solution ... but it looks like something along the lines of
hdparm -B 254 /dev/sda
or
hdparm -B 255 /dev/sda
might be worth considering (in /etc/rc.local).
larch: http://larch.berlios.de
Offline
Yeah, I've read about using hdparm, but as you said looks like there is no easy solution since there seems to be some risks using those hdparm setting.
Offline
about power usage :
I get around 8.1 to 8.6 without wifi and around 1 watt more with wifi !
some tips :
* you don't need to unload the ath5k module, just bring the wireless interface down :
ifconfig wmaster0 down
(I've added it to my rc.local, and also in the netcfg wireless profiles as a POST_DOWN option.)
* The Fn keys for volume and suspend are working with xbindkeys (I've edited the wiki)
Edit : with the latest kernel 2.6.27.10 the speakers are muting when a headset is plugged in. However the muting with "amixer sset Master toggle" does only work whith headset plugged in, otherwise nothing happens !?
I also find that the speaker volume is very low compared to windows...
Last edited by john_schaf (2008-12-24 17:34:09)
Offline
about power usage :
I get around 8.1 to 8.6 without wifi and around 1 watt more with wifi !some tips :
* you don't need to unload the ath5k module, just bring the wireless interface down :
ifconfig wmaster0 down
(I've added it to my rc.local, and also in the netcfg wireless profiles as a POST_DOWN option.)* The Fn keys for volume and suspend are working with xbindkeys (I've edited the wiki)
Edit : with the latest kernel 2.6.27.10 the speakers are muting when a headset is plugged in. However the muting with "amixer sset Master toggle" does only work whith headset plugged in, otherwise nothing happens !?
I also find that the speaker volume is very low compared to windows...
Thanks for editing the wiki.
Your power consumption looks similar to mine.
About the wifi. iwconfig gives me two devices wlan0 (currently working) and wmaster0. It seems that you use wmaster0 for that I get the message that it isnt' a wifi device.
Is there something going wrong?
Offline
That's normal. It because of the new wlan-stack. There will always be a virtual (right term?) device now.
Offline
Offline
about power usage :
* The Fn keys for volume and suspend are working with xbindkeys (I've edited the wiki)
For the wiki, did you use the "official" key combination for volume +/-?
I had to use different keycodes.
As for the patch that you find when you follow Cimis link: Do I have to set it manually or can I just wait because it will be merged into the official kernel?
Offline
For the wiki, did you use the "official" key combination for volume +/-?
I had to use different keycodes.
As for the patch that you find when you follow Cimis link: Do I have to set it manually or can I just wait because it will be merged into the official kernel?
* For the keycodes, I used the ones given by "xbindkeys -k"
I think they might be different if the keyboard is not the same (mine is a french one)
* For the patch, I have no clue wether it will make it to the official kernel, but I posted an archlinux bug report to ask for inclusion in arch-kernel...
http://bugs.archlinux.org/task/12606
Offline
concerning the "wake up twice" issue, I found that actually the sleep key "Fn+Esc" is recorded twice (look at /var/log/everything.log)
So one workaround for that is to edit the /etc/acpi/handler.sh to NOT handle the sleep button event.
Instead, use xbindkeys to bind it to pm-suspend.
It works perfectly (from X)
Other open issues are :
- low sound : sound volume is much lower as in windows. I've seen thant this issue has been reported on other systems too...
Offline
Last week I had to wake up three time from suspend.
I never used Fn+Esc, instead I configured acpi to pm-suspend when the lid is closed.
Do you only use pm-suspend or do you add --quirk-vbestate-restore as I do?
Yes, I have a german layout so this could explain the differences ;b
Furthermore thanks for posting a bug report.
Offline
Using something like this in handler.sh seems to take care of the need to resume multiple times after suspend:
button/lid)
if [[ `cat /proc/acpi/button/lid/LID0/state | awk '{print $2}'` = "closed" ]]
then
pm-suspend
fi
#echo "LID switched!">/dev/tty5
;;
Diff to the original handler.sh here -- I also created some additional empty events to prevent acpid from logging unnecessary warning messages. This setup does hibernate when Fn-Esc is pressed and suspend when the lid is closed.
If someone can confirm that this works for them as well maybe we can add it to the wiki...
I'm really impressed with this little machine, btw
Offline
So anybody noticing the HDD load unload cycle problem I mentioned earlier. I tried the 254 and 255 hdparm settings and had HDD temps +5c over XP and constant fans.
Is it just that you all are ignoring it and hoping it doesn't mess up your drives while it runs up?
Anybody want to post to cycles/hours.
It pains me to keep booting into xp when I've a nice minimal fluxbox waiting thats super fast, but it's the only choice atm it seems.
Offline
Personally I refuse to participate in the load cycle whatever paranoia. Modern hard drives are designed to spin down, and I want the hard drive on my laptop to spin down whenever it is not used (I'm using a 2 minute timeout -- configured via laptop-mode-tools). Mind you -- if your hd is spinning down as often as you say then it is misconfigured, somehow. This is from /etc/laptop-mode/laptop-mode.conf on my computer:
CONTROL_HD_IDLE_TIMEOUT=1
#
# Idle timeout values. (hdparm -S)
# Default is 2 hours on AC (NOLM_HD_IDLE_TIMEOUT_SECONDS=7200) and 20 seconds
# for battery and for AC with laptop mode on.
#
LM_AC_HD_IDLE_TIMEOUT_SECONDS=3600
LM_BATT_HD_IDLE_TIMEOUT_SECONDS=120
NOLM_HD_IDLE_TIMEOUT_SECONDS=7200
#
# Should laptop mode tools control the hard drive power management settings?
#
CONTROL_HD_POWERMGMT=1
#
# Power management for HD (hdparm -B values)
#
BATT_HD_POWERMGMT=250
LM_AC_HD_POWERMGMT=254
NOLM_AC_HD_POWERMGMT=254
This is what S.M.A.R.T. info analysis on my system looks like:
$ sudo hdsentinel /dev/sda
Hard Disk Sentinel for LINUX console 0.02 (c) 2008 info@hdsentinel.com
Start with -r [reportfile] to save data to report, -h for helpExamining hard disk configuration ...
HDD Device 0: /dev/sda
HDD Model ID : TOSHIBA MK1652GSX
HDD Serial No: Y84ZT07DT
HDD Revision : LV011E
HDD Size : 152628 MB
Interface : S-ATA II
Temperature : 29 °C
Health : 100 %
Performance : 100 %
Power on time: 1 days, 15 hours
Est. lifetime: more than 1000 days
That's good enough for me.
Offline
Thanks for posting your laptop-mode.conf file, I'll compare and adjust mine.
Offline
Hi fwojciec,
nice to see you in this thread joining the NC10 community
I patched my handler.sh with your diff file and it works like a charm - thanks for this.
So it seems that --quirk-vbestate-restore isn't needed (anymore?) since you can call pm-suspend in handler.sh without it.
I'm also not that concerned about spin ups. I think HDs will last long enough even with many spin ups and I rsync my files regularly.
What are your options in laptop-mode.conf for the "lost work settings" (this are the dirty_writeback_centisecs, aren't it?)
Currently I have
LM_BATT_MAX_LOST_WORK_SECONDS=15
LM_AC_MAX_LOST_WORK_SECONDS=5
but maybe this works against the idle settings?
Now a bit off-topic ;b
I saw your screenshots in the artwork section and I like it. I'm still playing around with my openbox setup. Currently I use a keybind for "ToggleFullscreen". It works for me but I'd like to see how it is if apps start without decorations. Is this in the rc.xml and may I see yours?
n obconf to prevent windows going over conky? I'm also interested in your .conkyrc. I guess you set a margin i Maybe I can see this, too (sorry if this is the wrong place)
Barghest
Offline
Hi Barghest -- thanks for getting the thread organized, it made installing Arch on this machine easy
Pm-suspend: I just have the quirk enabled by default:
$ cat /etc/hal/fdi/policy/99local-pm-utils-quirks.fdi
<?xml version="1.0" encoding="ISO-8859-1"?> <!-- -*- SGML -*- -->
<!-- Created by pm-utils -->
<deviceinfo version="0.2">
<device>
<match key="system.hardware.vendor" string="SAMSUNG ELECTRONICS CO., LTD.">
<match key="system.hardware.product" string="NC10">
<match key="system.firmware.version" string="02CA.MP00.20081015.KTW">
<match key="system.hardware.primary_video.vendor" int="0x8086">
<match key="system.hardware.primary_video.product" int="0x27ae">
<merge key="power_management.quirk.vbestate_restore" type="bool">true</merge>
</match>
</match>
</match>
</match>
</match>
</device>
</deviceinfo>
I have the same attitude with regard to laptop hd drives -- my laptops are never "storage devices" per se, they're constantly synced with my main machine, which is, in addition, automatically backed up to an external server, so even if the hard drive dies it's not a huge problem, just a temporary inconvenience. Replacing a laptop hd these days is cheap and relatively easy. I haven't actually had a hd failure yet (knock on wood), but if it happens it's not a big deal.
I'm still tweaking laptop-mode tools, and for now I just have the defaults for lost work seconds:
LM_BATT_MAX_LOST_WORK_SECONDS=600
LM_AC_MAX_LOST_WORK_SECONDS=360
I like to keep laptop-mode-tools close to the defaults. I'm unsure what exactly it is that this setting does (it does sound like the dirty writebacks thing) but, in either case, 10 minutes when on battery seems reasonable to me.
Openbox configuration -- This is what I like to use instead of toggle fullscreen action:
<keybind key="A-F11">
<action name="ToggleMaximizeFull"/>
<action name="ToggleDecorations"/>
</keybind>
The reason is that this doesn't cover up conky (single line at the top of the screen) so that I can always see time, battery charge and such. I also have a bunch of applications defined to start up maximized/undecorated by default (and then this key combination can be used to make them into regular decorated windows again). Anyways -- here's the full rc.xml
Offline
Thanks for posting your laptop-mode.conf file, I'll compare and adjust mine.
I did some experiments and it seems that the HD in NC10 is very similar to a different Toshiba HD I've used (and ended up replacing) on one of my old laptops -- it doesn't seem to care too much about settings set with hdparm and tends to do it's own thing instead. So I'm not completely sure whether my settings are any good
But I'll keep playing around with it and I'll post here if I discover something useful...
Offline
Thanks for your configs.
It's strange that pm-suspend works for me without setting a quirk.
Additional in Germany the NC10 is shipped either with a Hitachi or a Fujitsu HD (I have a Fujitsu), so no Toshiba drives here.
Offline
You are right -- the quirk is not needed. Although I am absolutely positive that I did need it at some stage, either suspend or hibernate wouldn't resume properly without it -- it was suspend I think.
Sometime along the way I upgraded the bios to the latest one from Samsung support site, and perhaps this explains why things are working correctly now.
Offline
This will help too http://www.phoronix.com/scan.php?page=a … nc10&num=1 an article about the NC10
Offline
My hd issue is ok now, I've it down to between 20-60 loads/hr which is fine by me. Thanks for the help. First time using arch and it's easily the best linux distro out there.
Offline