Force File Swap reduce RAM

HowTO
1 Image for example BlacHole 16.4 HDMI - Cec
1 USB with Swap 128 Megas

Uninstall ( If you have ) plugin E2 Free Memory
Place file attachament /etc ( attribute file = 644 )
Reboot Box

Flash is at 85% important .

This file force Data and inodes caches from RAM at Swap .
Start up RAM is 97 or 98% , please uses plugins , record , etc ... after 10 minutes reduce RAM .
Enjoy .
 

Attachments

  • 1.jpg
    1.jpg
    115 KB · Views: 100
  • 2.jpg
    2.jpg
    133.4 KB · Views: 86
  • CalamaresRellenos.zip
    225 bytes · Views: 101
  • screenshot.jpg
    screenshot.jpg
    54 KB · Views: 76

bh4me2

Vu+ Newbie
Interesting, default values:
vm.swappiness = 60
vm.vfs_cache_pressure = 100

New values:
vm.swappiness=100
vm.vfs_cache_pressure=50

Some reading:
Code:
http://www.linuxvox.com/2009/10/what-is-the-linux-kernel-parameter-vm-swappiness/
http://rudd-o.com/en/linux-and-free-software/tales-from-responsivenessland-why-linux-feels-slow-and-how-to-fix-that
Remove LSK via telnet:
ipkg remove enigma2-lsk-free-memory-v0.3
 

bh4me2

Vu+ Newbie
Some observation: the HDD does not go into standby anymore when using a swapfile, is this normal ? anyway to fix this ?
 
Interesting, default values:
vm.swappiness = 60
vm.vfs_cache_pressure = 100

New values:
vm.swappiness=100
vm.vfs_cache_pressure=50

IF you knows this so that not to write?

Ok.
You to try?
RAM ?
Ok
RAM 41 - 57 % ?

I have VU+DUO with motor tuner A Tuner B Fix to Astra 19 and TWO Tuner DTT , four 4 records online , It has used value by defect? NOT . NOT posible stable 4 records always . ¿¿??

Problem I to comment. Please admin to erase thread. Thanks.

P.D. Spanish = A mi los cojones no me los toca nadie , y menos un mascamierda .
 
Interesting, default values:
vm.swappiness = 60
vm.vfs_cache_pressure = 100

More info man ....

This parameters uses all distrubutions LINUX , example Debian ubuntu Mandrive etc ... BlackHole uses LINUX Kernel 2.6. XX , IF PC with Debian 2Gb RAM , ok , your verify file /etc ...
BlaskHole uses this defauls .
With Box Vu+DUO 141 Megas RAM

Please because not forcing Swap ? please because ?

Values testing ago 20 days ,

Your not to prove anything, total improvement, if you to prove well but please east nonboycott thread. Thanks
 

bh4me2

Vu+ Newbie
Huh?? the subtitles are missing :)
Seriously I can't understand what you are trying to say here.
I'm not boycotting you, only adding more information to your possible solution to prevent crashes with epg loads, such as what BH's default values are compared to yours and some info why some values are what they are. Also instruction were missing about removing LSK. If you provide a solution or improvement please try to be complete what people are suppose to do and provide info how to reverse it (rollback). Until a generic solution is made in BH we have to experiment with ideas like this one, the more ideas and the more people testing them then maybe we can have a solution for all BH users.

Anyway so far after a day with a smaller swapfile:
Code:
root@bm750:~# free
______total____used____free____shared___buffers
Mem:__141480__134716___6764____0_____1620
Swap:__129460__37252___92208
Total:__270940__171968___98972
 

bh4me2

Vu+ Newbie
4gb flashdrive (smallest one I could get) but a lot faster then my previous one:
Code:
root@bm750:~# free
              total         used         free       shared      buffers
  Mem:       141480       137812         3668            0          728
 Swap:       131064        50252        80812
Total:       272544       188064        84480
128mb swapfile but 50mb already used after 1 day !
(ps tnx mods for the code block hint)

bhstat1.jpg
 
Ok not problems , i have links for my idea force Swap .
Code:
http://debian-eleckttruss.blogspot.com/2007/04/como-configurar-swappiness-para-tener.html
http://lwn.net/Articles/83830/
http://yordanisp.blog.com.es/2010/08/06/optimizar-el-uso-de-memoria-ram-y-swap-en-linux-swappiness-9137452/
http://bredsaal.dk/autoadjusting-linux-vm-swappiness
http://linuxpoison.blogspot.com/2010/05/performance-tuning-linux-kernel-with.html
Etc ...

Enjoy . ;)
 

tobler

Vu+ User
i really don't understand all this discussion:

free
total used free shared buffers
Mem: 141476 94864 46612 0 828
Swap: 522104 0 522104
Total: 663580 94864 568716


46megs free and 0 swap used i think looks good
 

bh4me2

Vu+ Newbie
Tobler: do you auto update your EPG everyday via crossepg ? the issue is when epg is loaded the box may crash, this is one of the tests (a swapfile and forcing more use out of it with sysctl.conf) to see if we can find some sort of a solution. (maybe not but I'm trying it anyway)

Code:
root@bm750:~# free
              total         used         free       shared      buffers
  Mem:       141480       136060         5420            0         1460
 Swap:       131064        65484        65580
Total:       272544       201544        71000
Day 2, so far so good.
 

tobler

Vu+ User
no i don't use the crossepg, but the others weekly epgs present in the bh

what's the advantage of the crossepg one?
 

bh4me2

Vu+ Newbie
@tobler: Don't know, never got BH's internal epg working properly, the problem here is not BH's epg or crossepg but, as far as testing go, the moment epg data is loaded.

Code:
Ram usage
root@bm750:/# free
              total         used         free       shared      buffers
  Mem:       141480       134804         6676            0         1356
 Swap:       131064        64020        67044
Total:       272544       198824        73720

HDD Temp
root@bm750:/# hddtemp -n SATA:/dev/sda1
36

Flash use
root@bm750:/# df -h
Filesystem                Size      Used Available Use% Mounted on
/dev/root               118.0M     56.7M     61.3M  48% /
Day 3, EPG still processed automatically.
 

bh4me2

Vu+ Newbie
Day 4, I do notice that the box is getting a little bit sluggish when recording with the swap activities but no crashes yet, (cross)epg processed normal again. Also I noticed that the HDD is going into standby, maybe I was a bit early in mentioning that part wasn't working with a swapfile.
Code:
bm750 login: root
root@bm750:~# free
              total         used         free       shared      buffers
  Mem:       141480       133832         7648            0          992
 Swap:       131064        66728        64336
Total:       272544       200560        71984
root@bm750:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/root               120832     58072     62760  48% /
 

bh4me2

Vu+ Newbie
Some strange things started to happen, on the remote the OK button took me to the channel list and a Mute switched the box to standby mode, box still very sluggish, so reversed the vm settings to default, box rebooted. Test aborted.
This method can be used to prevent epg-box crashes but the box needs rebooting every 2-3 days which defeats the purpose.

Next: is to keep the swapfile, default vm settings, lsk not used, crossepg auto update daily.
Flash remained at all times at fixed 48% so assuming nothing changes in flash.

Note: removing LSK does not remove the bhcron job, use the bhcron menu (blue,blue) to remove that.
 

bh4me2

Vu+ Newbie
4 days uptime and still stable responsive box, swapuse is rising though.

Code:
root@bm750:~# free
              total         used         free       shared      buffers
  Mem:       141480       138208         3272            0          916
 Swap:       131064        30444       100620
Total:       272544       168652       103892
Lets see if I can make the weekend without a crash.
 

bh4me2

Vu+ Newbie
As assumed, amount of ram use is steadily rising over time, my guess is that epg data is not cleared properly, anyone any ideas how to clear epg space without a reboot ?? Still stable responsive box unlike the start of this thread with other swap settings, the default seems to keep vital code in ram.
Code:
root@bm750:~# uptime
 20:09:01 up 6 days, 20:31, load average: 0.00, 0.00, 0.00
root@bm750:~# free
              total         used         free       shared      buffers
  Mem:       141480       138404         3076            0          912
 Swap:       131064        52796        78268
Total:       272544       191200        81344
root@bm750:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/root               120832     58064     62768  48% /
 

bh4me2

Vu+ Newbie
Last update so far:
Code:
root@bm750:~# uptime
 13:41:42 up 8 days, 14:04, load average: 0.00, 0.00, 0.00
root@bm750:~# free
              total         used         free       shared      buffers
  Mem:       141480       138720         2760            0          452
 Swap:       131064        68764        62300
Total:       272544       207484        65060
root@bm750:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/root               120832     58068     62764  48% /
Epg still updated daily with crossepg, no crashes yet but ram use is rising! Just waiting now how long it will take before all ram is used and possibly a crash after that.
 

Tamaki

Administrator
Last update so far:
Code:
root@bm750:~# uptime
 13:41:42 up 8 days, 14:04, load average: 0.00, 0.00, 0.00
root@bm750:~# free
              total         used         free       shared      buffers
  Mem:       141480       138720         2760            0          452
 Swap:       131064        68764        62300
Total:       272544       207484        65060
root@bm750:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/root               120832     58068     62764  48% /
Epg still updated daily with crossepg, no crashes yet but ram use is rising! Just waiting now how long it will take before all ram is used and possibly a crash after that.
I think the problem is that everyday crossepg does a merge between old and new downloaded epg.
So ram use rises.
 
Top