insights get wheezy distro on iMac G4?[generally solved]

[huggybear]

I have already said that i don't understand sysctl.
That said:
I think it is the other way around than you said above.
- preload will help _if_ you got enough of RAM and load apps you use often.

I wouldn't know and Julian's guide doesn't comment on that. All I know is that on my old PII 233 MHz with 256 MB RAM preload helped a lot, so I presume it uses more swap.

- The entry for sysctl in it's config will start to use _swap_ more early (assuming you have set it for that purpose).

Hmm, yes and no. For vm.swappiness, the lower the value, the more RAM it uses rather than swap.
The greater the value, the more swap it uses as opposed to RAM.
Your value is 10, so it's lower than the Debian default (60) and therefore will prefer RAM to swap by a large margin compared to vanilla Debian. If you want swap rather than RAM, you'd set it to a higher value.

With vm cache pressure, it's the other way around -- the lower the value, the more swapping. But let me quote Julian on that:

We can choose a value between 0 and 100 where 0 means the kernel tries to keep everything in RAM and not cached to disk and 100 means it aggressively caches to disk to free RAM. The default in Debian is 60, which is OK but conservative. Laptop users should in any case use a low value to reduce writing to disk (because writing to disk negates benefits of power management and runs down your battery very quickly).

Code: Select all

 # echo 'vm.swappiness=20' >> /etc/sysctl.conf

The second change will affect the way the kernel frees caches of block devices vs filesystem entries. We'd prefer that when the kernel tries to free up cache it gets rid of cached block devices before cached filesystem entries, because having cached filesystem entries makes our filemanagers and other applications much more responsive than if they had to seek everything from the actual filesystem. This time we set a value for vm.vfs_cache_pressure, also in /etc/sysctl.conf. The default is 100, I don't know what the top limit might be but if you choose 0 your system may indulge in a horrible swapfest when handling large files. Try this:

Code: Select all

    # echo 'vm.vfs_cache_pressure=50' >> /etc/sysctl.conf

[huggybear]

The whole thing.

[nadir]

Your value is 10, so it's lower than the Debian default (60) and therefore will prefer RAM to swap by a large margin compared to vanilla Debian. If you want swap rather than RAM, you'd set it to a higher value.

You said it:
_Your_ value (which is mine).
I got 4 Gig of Ram, hence don't want to move it to swap. I only copied and pasted _ my_ entry as a search term to search for.
I didn't want to say those are good values, only quote the terms.
And i said:

(assuming you have set it for that purpose).

So: not "yes and no". Yes.

I should probably only have linked to the antix site, but it was more easy to find my entry (Then i added the antiX-link; afterwards).

I for one should remove the entry altogether.
Good that you fount the link/quote. Will help me to understand it.

[este_el_paz]

Gents: appreciate the clarifications . . . I also am now almost at the blithering idiot stage. Might get a chance to fiddle for a few minutes, see if I can log in to the Debian side of the HD w/o too many kernel panics . . . we're working with a raging 512MB of RAM in this unit, default swap, which I think is 1GB?? The simplest solution to getting passed the panics would be my goal, if just upgrading out of it would do it, or whatever the "Backports" option is . . . checking that out. Since I was chafing at the 4 hours install for the MintPPC11 and then this last time the Debian DVD got it down to 3+ . . . don't know if 6 hours to recompile the kernel is going to be easy to find, unless you can totally walk away from it after the first 15 minutes. Lots more Bat fun to come . . . .

e.e.p.

[nadir]

I was unclear:
With six hours i meant six hours of compiling (when the compiler runs and does his job). I watched TV or sleeped during that time (or whatever). I think i got 256 of Ram and (i think this is rather the problem) something like 700Mhz for CPU.
Else compiling a kernel can be very fast.
(configuring it, otoh, is really shit. Tons of options, most things i never heard of before... That is why i said: i would simply do a test-compilation _without_ making any changes to the configs during "make menuconfig". Just to get the feeling of it. Then make a few changes... and so on).

backports is here (with super easy instructions):
http://backports-master.debian.org/
Add the repo, update, and do a "apt-cache search linux-image".
If you are a lucky man apt-get installing a 2.6.38 kernel (which i assume is the backports one) will be added to yaboot. If you are more lucky that will solve your kernel-panic-problem.

What i would probably do now is search the web or ask at forums and mailing lists and IRC channels how on earth one can figure out _why_ a kernel panics. Else it's all shooting in the dark. But _if_ you know what the problem is...
I never had much problems with kernel panics, so ... no clue from here.

[nadir]

If you say when the kernel panics exactly (what is the last usual boot-output befor it panics?), it might be of help.
Not for me, in general (for others). Might be you said it above, and i missed it.

Try /etc/default/bootlogd from NO no YES
(did never help me much, but won't hurt. The log will be at /var/log/boot, readable only by root).

I would also unplug as much as possible (on my ppc that is close to impossible, but on my PC's i unplug all and everything, hard-disks, floppies, network cards, Ram-sticks, etc;; then replug them one after the other. Sometimes that helps).

[este_el_paz]

@nadir: funny, I was running a search of "backports" on the forum search bar and your latest post was at the top of the list . . . ??? Hows that for search "foo"??? Anyway, thanks for your able assist; I am booted into a GNOME/openbox DE right now, had no kernel panics getting into it, for some reason I like GNOME, doesn't seem so cramped. Anyway, following your post comment about where to find the system timer info:

f) for the record: /lib/modules/***kernel-version***

I went to the directory tree to see if I could GUI my way to the file where the sytem timer data would be found in the lib/modules/3.1.0-1-powerpc directory/folder but after rifling and opening a number of, I call them folders, . . . nothing made itself too obvious. Still thinking about this T-mode FW nano/edit special ops attack . . . any thoughts about which exact file would contain the system timer data??? I guess I could try to search the computer with "system timer"?? still getting used to the different layouts in the various DE's which seem to have access to different apps than each other?? Anyway, at least there were no panics to boot into the Deb side this time . . . tried to log into what I thought would be a plain "openbox" session which showed up in the login manager . . . but that was an empty screen, had to shut down--expecting to crash on relaunch, but all went well!!!

e.e.p.

[nadir]

Ups...
I think you said you are running testing, right?
In that case speaking of "backports-kernels" is utter nonsense (my fault).
If one run's stabel one can use stable-backports to install newer packages (say a kernel).
If one runs testing one can run a mixed testing/unstable system.
For that you would need to edit /etc/apt/preferences and set testing to be the default.
I think in case of the kernel you can also add the sid (unstable) repos to your sources list,
update, install the kernel, and comment the sid-line in /etc/apt/sources list again. -> Don't forget to run "apt-get update" again, else the next "apt-get upgrade" will happily upgrade your system to unstable/sid (not the end of the world, but if one doesn't want it...)
(or download the deb's and install them manually).
Else i would not do something like that, but in case of the kernel i think it's ok (I said I think, not i know).

Bout your actual news: Good. If it was me i would be happy with it, pray during each boot that it won't panic, and call it a day... Not saying that is the perfect approach (ignoring problems...)

[este_el_paz]

Gents: I'm just trying to clean this up as per Mr-Bisquit's request that I post what I did to get the wheezy system (more or less) working. First thing I had to do was repair the apparently broken ethernet connection, which I did thru the Terminal following instructions for trying various commands from the MintPPC.org forum archives. After I did that I tried a few things to get the "nv" driver installed but it kept coming up "file not found" . . . . I think I had already done mr-bisquit's suggestion of adjusting the "vert and horz" refresh rates as per his post in this forum, but without immediate success. Reading further over on the MintPPC forum where there was some discussion about getting the video drivers to work on G5 iMac's I posted a question there about my problems even getting the "nv" driver through the Wheezy "apt-get update/upgrade" command and then running "apt-get install xserver-xorg-video-nv" . . .and the moderator replied that "nv is not available in the wheezy repository" with the following suggested commands--which I did. I had already done the "modprobe" of nouveau so at the end of the install nv the screen flashed pink and then the Debian background image came up. I also had already adjusted the X11/xorg.conf file to the "nv" driver as well. I followed HB's advice to use the "lightdm" and I finally also got a basic Xfce DE downloaded, so I have a few DE to boot into as the mood suits. So that part is generally solved, the graphics are good enough . . . it's there for a rainy southern california day to mess with . . . it's just now these kernel panics that have to be passed through to boot into the Debian system AND the fact that apparently nv doesn't let the system recover from "suspend" . . . necessitating a restart . . . and then back into the kernel panics mode. Among other possible sources of these panics, the system clock seems to reset itself to the wrong time, even though the date & time map shows my correct location. I'll possibly start a new thread on the kernel panics thing in a short minute or three. Thanks for the help here.

e.e.p.

to get the nv package installed and working properly, do the following:
login as root with root password
Code:
nano /etc/apt/sources.list

scroll down to the bottom and add this line:
Code:
deb http://mintppc.org/repository katya main

CTRL-X and "y" to save
Code:
apt update
apt install xserver-xorg-video-nv
echo "blacklist nouveau" > /etc/modprobe.d/blacklist-nouveau.conf
nano /etc/X11/xorg.conf

paste the following into it:
Code:
Section "Device"
Identifier "nvidia"
Driver "nv"
EndSection

Section "Extensions"
Option "Composite" "enable"
EndSection

CTRL-X and "y" to save
Code:
reboot