Re: [arch-general] Munin and MySQL
On Mon, Sep 17, 2012 at 2:40 AM, Martín Cigorraga m...@archlinux.us wrote:
Hi all, anyone proficient at Munin?
Im no pro, but recently did a install again.
The talk of one of our guest speakers at the local SFD that took place
yesterday was about monitoring everything with Munin, a software I didn't
knew but which I liked at first sight.
Once at home the very first thing I did after take the laptop out of my
backpack was install Munin, of course, and alongside this I found some
warning and error messages that aren't very informative so before go and
dive on the net trying to gather information about them I decided to first
try luck in the list.
Whenever I run:
# munin-node-configure --shell | sh
I got this:
# The following plugins caused errors:
# mysql_:
# Non-zero exit during autoconf (2)
# ntp_states:
# Non-zero exit during autoconf (2)
# proc:
# In family 'auto' but doesn't have 'autoconf' capability
Now the munin-node-configure --suggest command shows:
| no | no
for all the three plugins. Following the wiki suggestion:
# munin-run {plugin}
I got a:
# Unknown service '{plugin}'
for all them three.
Are you 'literally' running munin-run {plugin} or are you specifying
the plugin?
What does the following command output? eg:
[somedude@somebox plugins] # munin-run --debug mysql_threads
# Setting /rgid/ruid/ to /997/99/
# Setting /egid/euid/ to /997 997/99/
# Setting up environment
# About to run '/etc/munin/plugins/mysql_threads'
threads.value 1
--
Regards
L.
Sent using electronic mail ツ
Re: [arch-general] Munin and MySQL
2012/9/17 Martín Cigorraga m...@archlinux.us: Whenever I run: # munin-node-configure --shell | sh I got this: # The following plugins caused errors: # mysql_: # Non-zero exit during autoconf (2) # ntp_states: # Non-zero exit during autoconf (2) # proc: # In family 'auto' but doesn't have 'autoconf' capability Try: /usr/lib/munin/plugins/mysql_ suggest If in doubt, read plugin source. This one needs some extra perl packages (it will show helpful error message if dependencies are missing), configuration in /etc/munin/plugin-conf.d/ (in existing or new file, no difference) and meaningful symlink name. If you want to install that plugin manually try something like this: ln -s /usr/lib/munin/plugins/mysql_ /etc/munin/plugins/mysql_$GRAPH Where $GRAPH is one from: (bin_relay_log, commands, connections, files_tables, innodb_bpool, innodb_bpool_act, innodb_insert_buf, innodb_io, innodb_io_pend, innodb_log, innodb_rows, innodb_semaphores, innodb_tnx, myisam_indexes, network_traffic, qcache, qcache_mem, replication, select_types, slow, sorts, table_locks, tmp_tables). -- Krzysztof Warzecha
Re: [arch-general] Proprietary nvidia driver for different kernels
On Sun, 2012-09-16 at 21:30 +0200, Joakim Hernberg wrote: Have you tried nouveau lately? It has improved a lot lately... Thank you :) indeed it works with the kernel-rt. I still have to test it a little bit longer and to check if audio production will be ok too, but until now it's ok for averaged desktop usage. Unfortunately I can't install nouveau-dri, since this conflicts with libgl and nvidia-utils and I'll use the proprietary driver with the standard kernel. Only drawback, after startup and when starting the Xfce session, the monitor relay is switching very often. The advantage is, that I don't need to readjust the monitor when switching between nouveau and nvidia. In the past nouveau only was a PITA, but now everything seems to be fine. Regards, Ralf
[arch-general] sleep + lid events + kde = new problem
Using fully updated testing repo - starting 'recently' - I notice that lid close no longer sleeps my laptop. Kde power is configured to sleep on lid close. If i click the menu and choose sleep manually - it sleeps fine - and subsequently wakes fine on lid-open. I can confirm with this little shell loop (left running and then close lid and reopen) that acpi shows lid is indeed closed: while [[ 1 == 1 ]] do cat /proc/acpi/button/lid/*/state sleep 1 done state: open state: open state: open state: open state: closed state: closed state: closed state: closed state: open state: open So, it -seems- as if lid-event is not being passed along for some reason to the (kde) power management - or not being listened to. I have not yet switched to systemd - and I'm not sure which process is responsible for such events. It still happens with 3.5.4 kernel if that makes any difference. It is a W520 lenovo laptop. Anyone else notice similar - or can offer suggestions? Thanks. gene/
[arch-general] HARDWARECLOCK and systemd
I'm moving to systemd (finally!) and I'm stuck at the HARDWARECLOCK section of rc.conf wiki. How should be configured /etc/adjtime ? I can't find what should be inside it. Also there say: If this value is not set, then the value stored by hwclock in /var/lib/hwclock/adjtime [...] but I don't have any /var/lib/hwclock directory. I would appreciate if anyone who already did the jump can share his experiences, thanks!
Re: [arch-general] sleep + lid events + kde = new problem
On Mon, Sep 17, 2012 at 3:11 PM, Genes MailLists li...@sapience.com wrote: Using fully updated testing repo - starting 'recently' - I notice that lid close no longer sleeps my laptop. Kde power is configured to sleep on lid close. If i click the menu and choose sleep manually - it sleeps fine - and subsequently wakes fine on lid-open. I can confirm with this little shell loop (left running and then close lid and reopen) that acpi shows lid is indeed closed: while [[ 1 == 1 ]] do cat /proc/acpi/button/lid/*/state sleep 1 done state: open state: open state: open state: open state: closed state: closed state: closed state: closed state: open state: open So, it -seems- as if lid-event is not being passed along for some reason to the (kde) power management - or not being listened to. I have not yet switched to systemd - and I'm not sure which process is responsible for such events. It still happens with 3.5.4 kernel if that makes any difference. It is a W520 lenovo laptop. Anyone else notice similar - or can offer suggestions? Thanks. Does running acpi_listen and closing then reopening the lid gives an indication that the lid open/close is being seen - just to check if there is any switch failure? -- mike c
Re: [arch-general] HARDWARECLOCK and systemd
On 09/17/2012 08:53 AM, Martín Cigorraga wrote: I'm moving to systemd (finally!) and I'm stuck at the HARDWARECLOCK section of rc.conf wiki. How should be configured /etc/adjtime ? I can't find what should be inside it. Also there say: If this value is not set, then the value stored by hwclock in /var/lib/hwclock/adjtime [...] but I don't have any /var/lib/hwclock directory. I would appreciate if anyone who already did the jump can share his experiences, thanks! How did you have it set with initscripts? If it was blank or UTC, then you don't need to do anything to migrate to systemd. I can't speak for localtime
Re: [arch-general] HARDWARECLOCK and systemd
Am 17.09.2012 16:53, schrieb Martín Cigorraga: I'm moving to systemd (finally!) and I'm stuck at the HARDWARECLOCK section of rc.conf wiki. How should be configured /etc/adjtime ? I can't find what should be inside it. Also there say: If this value is not set, then the value stored by hwclock in /var/lib/hwclock/adjtime [...] but I don't have any /var/lib/hwclock directory. I would appreciate if anyone who already did the jump can share his experiences, thanks! The file is now in /etc/adjtime. It should already be there. If not, you set the time and run hwclock --systohc --utc or hwclock --systohc --localtime This creates the right configuration. signature.asc Description: OpenPGP digital signature
[arch-general] /etc/tmpfiles.d
[...]However, tmpfiles may also be used to write values into certain files on boot. For example, if you use /etc/rc.local to disable wakeup from USB devices with echo USBE /proc/acpi/wakeup, you may use the following tmpfile instead: /etc/tmpfiles.d/disable-usb-wake.conf w /proc/acpi/wakeup - - - - USBE The tmpfiles method is recommended in this case since systemd doesn't actually support /etc/rc.local. Does that means that I need to move all the content from /etc/rc.local to /etc/tmpfiles.d? For example this is my actual /etc/rc.local: ~ $ cat /etc/rc.local #!/bin/bash # # /etc/rc.local: Local multi-user startup script. # #modprobe radeon # added by hybrid-video-ati-intel install script #echo IGD /sys/kernel/debug/vgaswitcheroo/switch # added by hybrid-video-ati-intel install script echo OFF /sys/kernel/debug/vgaswitcheroo/switch # completely deactivate radeon ## ATi # Source: https://wiki.archlinux.org/index.php/ATI#Performance_tuning echo low /sys/class/drm/card0/device/power_profile #echo profile /sys/class/drm/card0/device/power_method echo dynpm /sys/class/drm/card0/device/power_method echo OFF /sys/kernel/debug/vgaswitcheroo/switch # CPUFREQ for i in 0 1 2 3; do cpufreq-set -c $i -g powersave; done ## sets powersave cpufreq governor for all CPU cores #echo -n 90 /sys/devices/system/cpu/cpufreq/ondemand/up_threshold echo -n 20 /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor # Prepare the system for Wake-on-Lan /usr/sbin/ethtool -s eth0 wol pg # Activate laptop_mode echo 5 /proc/sys/vm/laptop_mode # Performance tweaks for USB drivers under KDE SC echo madvise /sys/kernel/mm/transparent_hugepage/enabled echo madvise /sys/kernel/mm/transparent_hugepage/defrag echo 0 /sys/kernel/mm/transparent_hugepage/khugepaged/defrag If this is the case, how do you guys would convert the FOR loop!?
Re: [arch-general] /etc/tmpfiles.d
Hello. Does that means that I need to move all the content from /etc/rc.local to /etc/tmpfiles.d? For example this is my actual /etc/rc.local: Only if you no longer boot with initscripts. # CPUFREQ for i in 0 1 2 3; do cpufreq-set -c $i -g powersave; done ## sets powersave cpufreq governor for all CPU cores #echo -n 90 /sys/devices/system/cpu/cpufreq/ondemand/up_threshold echo -n 20 /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor (...) If this is the case, how do you guys would convert the FOR loop!? I think that tempfiles can not execute commands, only create/write to files. If you want to run some commands, you should probably create some custom unit files. Regards, Chris Sakalis
Re: [arch-general] Proprietary nvidia driver for different kernels
On Mon, 17 Sep 2012 15:27:04 +0200 Ralf Mardorf ralf.mard...@alice-dsl.net wrote: Unfortunately I can't install nouveau-dri, since this conflicts with libgl and nvidia-utils and I'll use the proprietary driver with the standard kernel. If I were you, I'd try nouveau all the way. only reason not to would be gaming. One thing I still have to check out though is if the support for vblank is working, otherwise it might have some tearing when watching video. But I agree, it's a pain switching between nvidia and nouveau, but I'd be inclined to blame nvidia itself for that...:) In any case it's nice to see that a opensource alternative to the closed nividia driver is available and working better all the time! -- Joakim
Re: [arch-general] /etc/tmpfiles.d
On 09/17/2012 09:40 AM, Martín Cigorraga wrote: [...]However, tmpfiles may also be used to write values into certain files on boot. For example, if you use /etc/rc.local to disable wakeup from USB devices with echo USBE /proc/acpi/wakeup, you may use the following tmpfile instead: /etc/tmpfiles.d/disable-usb-wake.conf w /proc/acpi/wakeup - - - - USBE The tmpfiles method is recommended in this case since systemd doesn't actually support /etc/rc.local. Does that means that I need to move all the content from /etc/rc.local to /etc/tmpfiles.d? For example this is my actual /etc/rc.local: ~ $ cat /etc/rc.local #!/bin/bash # # /etc/rc.local: Local multi-user startup script. # #modprobe radeon # added by hybrid-video-ati-intel install script #echo IGD /sys/kernel/debug/vgaswitcheroo/switch # added by hybrid-video-ati-intel install script echo OFF /sys/kernel/debug/vgaswitcheroo/switch # completely deactivate radeon ## ATi # Source: https://wiki.archlinux.org/index.php/ATI#Performance_tuning echo low /sys/class/drm/card0/device/power_profile #echo profile /sys/class/drm/card0/device/power_method echo dynpm /sys/class/drm/card0/device/power_method echo OFF /sys/kernel/debug/vgaswitcheroo/switch # CPUFREQ for i in 0 1 2 3; do cpufreq-set -c $i -g powersave; done ## sets powersave cpufreq governor for all CPU cores #echo -n 90 /sys/devices/system/cpu/cpufreq/ondemand/up_threshold echo -n 20 /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor # Prepare the system for Wake-on-Lan /usr/sbin/ethtool -s eth0 wol pg # Activate laptop_mode echo 5 /proc/sys/vm/laptop_mode # Performance tweaks for USB drivers under KDE SC echo madvise /sys/kernel/mm/transparent_hugepage/enabled echo madvise /sys/kernel/mm/transparent_hugepage/defrag echo 0 /sys/kernel/mm/transparent_hugepage/khugepaged/defrag If this is the case, how do you guys would convert the FOR loop!? For ethtool, just create a separate service that executes that command. Everything else you do is writing to /sys, so you can have one giant tmpfiles.d file. For the for loop: w /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu1/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu2/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu3/cpufreq/scaling_governor - - - - powersave Also, I don't think it's an error if the file doesn't exist, so you can just do cpu0..cpu16 or whatever if you feel like.
Re: [arch-general] /etc/tmpfiles.d
Am 17.09.2012 17:40, schrieb Martín Cigorraga: # CPUFREQ for i in 0 1 2 3; do cpufreq-set -c $i -g powersave; done ## sets powersave cpufreq governor for all CPU cores #echo -n 90 /sys/devices/system/cpu/cpufreq/ondemand/up_threshold echo -n 20 /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor Slightly off-topic, but you should really read this: http://mjg59.livejournal.com/88608.html signature.asc Description: OpenPGP digital signature
Re: [arch-general] sleep + lid events + kde = new problem
I can confirm this, I started seeing this behavior about 2 months ago. But in my case, it sometimes work, and sometimes doesn't whereas it used to work flawlessly before. --Héctor Acosta On Mon, Sep 17, 2012 at 10:13 AM, mike cloaked mike.cloa...@gmail.com wrote: On Mon, Sep 17, 2012 at 3:11 PM, Genes MailLists li...@sapience.com wrote: Using fully updated testing repo - starting 'recently' - I notice that lid close no longer sleeps my laptop. Kde power is configured to sleep on lid close. If i click the menu and choose sleep manually - it sleeps fine - and subsequently wakes fine on lid-open. I can confirm with this little shell loop (left running and then close lid and reopen) that acpi shows lid is indeed closed: while [[ 1 == 1 ]] do cat /proc/acpi/button/lid/*/state sleep 1 done state: open state: open state: open state: open state: closed state: closed state: closed state: closed state: open state: open So, it -seems- as if lid-event is not being passed along for some reason to the (kde) power management - or not being listened to. I have not yet switched to systemd - and I'm not sure which process is responsible for such events. It still happens with 3.5.4 kernel if that makes any difference. It is a W520 lenovo laptop. Anyone else notice similar - or can offer suggestions? Thanks. Does running acpi_listen and closing then reopening the lid gives an indication that the lid open/close is being seen - just to check if there is any switch failure? -- mike c
Re: [arch-general] Proprietary nvidia driver for different kernels
I don't play computer games. Perhaps I'll render 3D videos or use Google Earth. I'll give it a try and remove the proprietary driver. Watching videos and editing videos must be possible for my needs too. I was pissed off when the old nouveau driver didn't work for my machine and some distros already dropped the nv driver. If 3D should work with the kernel-rt, I don't need the standard kernel anymore, resp. perhaps I'll keep a second kernel for emergency cases. Hopefully the nouveau driver won't cause real-time issues. I guess I'll run a jack midi latency test later or tomorrow. In any case it's nice to see that a opensource alternative to the closed nividia driver is available and working better all the time! Unfortunately we only can use it until 21 December 2012 ;D. Regards, Ralf
Re: [arch-general] Proprietary nvidia driver for different kernels
On 17 Sep 2012 18:46, Ralf Mardorf ralf.mard...@alice-dsl.net wrote: I don't play computer games. In any case it's nice to see that a opensource alternative to the closed nividia driver is available and working better all the time! Unfortunately we only can use it until 21 December 2012 ;D. Unfortunately I have to use nvidia drivers for CUDA programming.. I might get a specific card just for the CUDA stuff and still load the nouveau for xorg. I don't think its possible with just one GPU, is it? Switching from nvidia to nouveau using the scripts in the wiki for now. Regards, Ralf
Re: [arch-general] sleep + lid events + kde = new problem
On 09/17/2012 11:13 AM, mike cloaked wrote: Does running acpi_listen and closing then reopening the lid gives an indication that the lid open/close is being seen - just to check if there is any switch failure? Tried this - I note that acpid was not running and it is needed for acpi_listen - so I started it - closed lid and re-opened. (laptop did not sleep after starting acpid either). acpi_listen then reports: % acpi_listen button/lid LID close button/lid LID open gene/
Re: [arch-general] Proprietary nvidia driver for different kernels
Joakim Hernberg jbh at alchemy.lu writes: But I agree, it's a pain switching between nvidia and nouveau, but I'd be inclined to blame nvidia itself for that...:) Not their fault. But nevertheless, they are the ones trying to fix it: http://lists.freedesktop.org/archives/mesa-dev/2012-September/027295.html
Re: [arch-general] Proprietary nvidia driver for different kernels
On Mon, 2012-09-17 at 19:02 +0100, Leonidas Spyropoulos wrote: On 17 Sep 2012 18:46, Ralf Mardorf ralf.mard...@alice-dsl.net wrote: I don't play computer games. In any case it's nice to see that a opensource alternative to the closed nividia driver is available and working better all the time! Unfortunately we only can use it until 21 December 2012 ;D. Unfortunately I have to use nvidia drivers for CUDA programming.. I might get a specific card just for the CUDA stuff and still load the nouveau for xorg. I don't think its possible with just one GPU, is it? Switching from nvidia to nouveau using the scripts in the wiki for now. I don't know. Btw. I was switching between nv and nvidia for several distros using a very primitive script: $ cat /etc/rc.d/69switch_xorg.conf #! /bin/sh # /etc/rcS.d/Switch_xorg_conf rm /etc/X11/xorg.conf case $(uname -r) in *rt*) cp /etc/X11/xorg.conf.nv /etc/X11/xorg.conf ;; *) cp /etc/X11/xorg.conf.nvidia /etc/X11/xorg.conf ;; esac At the moment nv is for nouveau. I didn't know those scrips: https://wiki.archlinux.org/index.php/NVIDIA#Switching_between_nvidia_and_nouveau_drivers If I should use those scripts, I guess I would replace pacman -S, by pacman -U to install always the versions I know as being stable for my needs. Perhaps I only need nouveau, have no time to test it now, I'll watch The Simpsons at the moment ... commercial break finished ... Regards, Ralf Regards, Ralf
Re: [arch-general] / mounted ro after update
On 09/15/2012 11:24 AM, P .NIKOLIC wrote: Of course noone else will offer any assistance will they EH! Pete . Pete, You know me, I'm never quite to the ire of many, but I haven't a clue what you are dealing with here... Let us know what it ultimately turns out to be and if it is an issue with util-linux on Arch's end, file a report so it gets fixed before it bites me :) -- David C. Rankin, J.D.,P.E.
[arch-general] Strange pacman problem with conflicting dependencies
Hi On a machine with a less frequent updating schedule (~once a month) I receive the following when doing pacman -Suy -8- # pacman -Suy :: Synchronizing package databases... testing is up to date core is up to date extra is up to date community-testing is up to date community is up to date multilib-testing is up to date multilib is up to date archlinuxfr is up to date repo-ck is up to date :: Starting full system upgrade... warning: ffmpeg: ignoring package upgrade (20120509-1 = 1:0.11.1-1) :: Replace virtualbox-additions with community-testing/virtualbox-iso- additions? [Y/n] :: Replace virtualbox-modules with community-testing/virtualbox-host-modules? [Y/n] warning: x264: ignoring package upgrade (20120204-1 = 20120705-1) resolving dependencies... looking for inter-conflicts... error: unresolvable package conflicts detected error: failed to prepare transaction (conflicting dependencies) :: virtualbox-host-modules and virtualbox-modules are in conflict -8- I didn't have any problems with the virtualbox replacements on another machine with daily updates. If necessary the debug pacman log may be found at http://pastebin.com/9nh5tp9t Thanks. -- Arthur Titeica
Re: [arch-general] / mounted ro after update
I had a problem of this sort recently, due to my root partition being full, so i went ahead and performed a: # pacman -Sc Before trying to update once again and things were rolling as normal back again, hope it helps. German C.
Re: [arch-general] /etc/tmpfiles.d
On 09/17/12 at 10:31am, Matthew Monaco wrote: On 09/17/2012 09:40 AM, Mart?n Cigorraga wrote: [...]However, tmpfiles may also be used to write values into certain files on boot. For example, if you use /etc/rc.local to disable wakeup from USB devices with echo USBE /proc/acpi/wakeup, you may use the following tmpfile instead: /etc/tmpfiles.d/disable-usb-wake.conf w /proc/acpi/wakeup - - - - USBE The tmpfiles method is recommended in this case since systemd doesn't actually support /etc/rc.local. Does that means that I need to move all the content from /etc/rc.local to /etc/tmpfiles.d? For example this is my actual /etc/rc.local: ~ $ cat /etc/rc.local #!/bin/bash # # /etc/rc.local: Local multi-user startup script. # #modprobe radeon # added by hybrid-video-ati-intel install script #echo IGD /sys/kernel/debug/vgaswitcheroo/switch # added by hybrid-video-ati-intel install script echo OFF /sys/kernel/debug/vgaswitcheroo/switch # completely deactivate radeon ## ATi # Source: https://wiki.archlinux.org/index.php/ATI#Performance_tuning echo low /sys/class/drm/card0/device/power_profile #echo profile /sys/class/drm/card0/device/power_method echo dynpm /sys/class/drm/card0/device/power_method echo OFF /sys/kernel/debug/vgaswitcheroo/switch # CPUFREQ for i in 0 1 2 3; do cpufreq-set -c $i -g powersave; done ## sets powersave cpufreq governor for all CPU cores #echo -n 90 /sys/devices/system/cpu/cpufreq/ondemand/up_threshold echo -n 20 /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor # Prepare the system for Wake-on-Lan /usr/sbin/ethtool -s eth0 wol pg # Activate laptop_mode echo 5 /proc/sys/vm/laptop_mode # Performance tweaks for USB drivers under KDE SC echo madvise /sys/kernel/mm/transparent_hugepage/enabled echo madvise /sys/kernel/mm/transparent_hugepage/defrag echo 0 /sys/kernel/mm/transparent_hugepage/khugepaged/defrag If this is the case, how do you guys would convert the FOR loop!? For ethtool, just create a separate service that executes that command. Everything else you do is writing to /sys, so you can have one giant tmpfiles.d file. For the for loop: w /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu1/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu2/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu3/cpufreq/scaling_governor - - - - powersave Also, I don't think it's an error if the file doesn't exist, so you can just do cpu0..cpu16 or whatever if you feel like. I am fairly certain that tmpfiles.d understands *, so you could probably get away with one line for something like that. -- Curtis Shimamoto sugar.and.scru...@gmail.com
Re: [arch-general] /etc/tmpfiles.d
On 09/17/2012 06:44 PM, Curtis Shimamoto wrote: On 09/17/12 at 10:31am, Matthew Monaco wrote: On 09/17/2012 09:40 AM, Mart?n Cigorraga wrote: [...]However, tmpfiles may also be used to write values into certain files on boot. For example, if you use /etc/rc.local to disable wakeup from USB devices with echo USBE /proc/acpi/wakeup, you may use the following tmpfile instead: /etc/tmpfiles.d/disable-usb-wake.conf w /proc/acpi/wakeup - - - - USBE The tmpfiles method is recommended in this case since systemd doesn't actually support /etc/rc.local. Does that means that I need to move all the content from /etc/rc.local to /etc/tmpfiles.d? For example this is my actual /etc/rc.local: ~ $ cat /etc/rc.local #!/bin/bash # # /etc/rc.local: Local multi-user startup script. # #modprobe radeon # added by hybrid-video-ati-intel install script #echo IGD /sys/kernel/debug/vgaswitcheroo/switch # added by hybrid-video-ati-intel install script echo OFF /sys/kernel/debug/vgaswitcheroo/switch # completely deactivate radeon ## ATi # Source: https://wiki.archlinux.org/index.php/ATI#Performance_tuning echo low /sys/class/drm/card0/device/power_profile #echo profile /sys/class/drm/card0/device/power_method echo dynpm /sys/class/drm/card0/device/power_method echo OFF /sys/kernel/debug/vgaswitcheroo/switch # CPUFREQ for i in 0 1 2 3; do cpufreq-set -c $i -g powersave; done ## sets powersave cpufreq governor for all CPU cores #echo -n 90 /sys/devices/system/cpu/cpufreq/ondemand/up_threshold echo -n 20 /sys/devices/system/cpu/cpufreq/ondemand/sampling_down_factor # Prepare the system for Wake-on-Lan /usr/sbin/ethtool -s eth0 wol pg # Activate laptop_mode echo 5 /proc/sys/vm/laptop_mode # Performance tweaks for USB drivers under KDE SC echo madvise /sys/kernel/mm/transparent_hugepage/enabled echo madvise /sys/kernel/mm/transparent_hugepage/defrag echo 0 /sys/kernel/mm/transparent_hugepage/khugepaged/defrag If this is the case, how do you guys would convert the FOR loop!? For ethtool, just create a separate service that executes that command. Everything else you do is writing to /sys, so you can have one giant tmpfiles.d file. For the for loop: w /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu1/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu2/cpufreq/scaling_governor - - - - powersave w /sys/devices/system/cpu/cpu3/cpufreq/scaling_governor - - - - powersave Also, I don't think it's an error if the file doesn't exist, so you can just do cpu0..cpu16 or whatever if you feel like. I am fairly certain that tmpfiles.d understands *, so you could probably get away with one line for something like that. It doesn't for the 'w' type. Globbing is specified explicitly where it's supported. I don't know if that was an explicit design decision though. You might have yourself an easy patch if you want to contribute.
Re: [arch-general] sleep + lid events + kde = new problem
Seems other events have stopped being processed as well - e.g. unplugging a/c and going to battery - laptop beeps - but screen no longer dims - Again - I can dim screen by hand by doing something like this: echo '11' /sys/class/backlight/acpi_video0/brightness This has been flaky a while but usually worked at least some of the time - now its just not working period. I wonder if udev responsible for this? Looks like I upgraded systemd at the end of last month to 189-4 ... seems unlikely it took me till last weekend to notice this .. but I can be slow at times :-) kde had a bunch of updates sep 4 (4.9.0 to 4.9.1).. again a week or so back If anyone has suggestions how to get these working again would be appreciated .. not fatal as I can do everything by hand but since things used to work be ncie to get them working again. gene
