My unhappy adventures with Linux installation


Ah, nouveaux.  There's been many announcements that it's been fixed since I last used it, but even more reports that actually it's still rather broken on specific cards.  I've had far more success using the closed binary-blob nvidia drivers on machines with those cards installed in the past.

It does sound like it's not quite as broken as it was when I last used it on an Nvidia Quadro card.  I don't think I even got a picture once it entered graphical mode on that build.
 
Before the distupgrade I was using "nv" installed manually because the one that Debian installed was freezing at the boot, but the 3D acceleration was not working, but after the upgrade I just see nouveau everywhere and I didn't set anything, but GLX things now works perfectly
 
Ah, nouveaux. There's been many announcements that it's been fixed since I last used it, but even more reports that actually it's still rather broken on specific cards.  I've had far more success using the closed binary-blob nvidia drivers on machines with those cards installed in the past.


 


It does sound like it's not quite as broken as it was when I last used it on an Nvidia Quadro card.  I don't think I even got a picture once it entered graphical mode on that build.
My experience is that it is highly dependent on the graphics card which driver works better. On my home PC both drivers work acceptable, altough the binary driver has better 3D performance, of course. On my work PC the binary driver is broken and nouveau works fine (yes, even suspend works :D ). Both machines have a NVIDIA PCI card (not the same).

So from your last post i gathered that you're using Debian? Which kernel version do you have? Did you upgrade to Debian 8?
 
Last edited by a moderator:
So from your last post i gathered that you're using Debian? Which kernel version do you have? Did you upgrade to Debian 8?
Until a few days ago I had Wheezy 7 LXDE, then I upgraded to Jessi 8, and the shutdown problems started

Code:
Linux powerdebian 3.16.0-4-686-pae #1 SMP Debian 3.16.7-ckt9-3~deb8u1 (2015-04-24) i686 GNU/Linux
 
I found some kind of solution for the shutdown issue... if I click the "shutdown" option from the LXDE menu AND suddenly I switch to a linux console, the PC will always turn off of the power correctly.

So it is really something related to what it is graphically drawing during the halt command...

But I have now another issue... that it was corrected after the upgrade, but now suddenly reappeared...

when I switch to 640x480 resolution, seems like the left and right borders virtually overflow out of the screen but are then drawn inside the screen mirrored... if I move the mouse pointer on the borders I can see the pointer both on the left and on the right...

The main problem is that many Wine games runs fullscreen with that resolution... yesterday I played (finally) Baldur's Gate GOG at fullscreen, with the original GOG configuration without issues, but today magically I have the mirrored resolution issue...

With xrandr I tried every resolution listed, and all are working right, the only one wrong is the 640x480...

What parameter should I change (at least in xrandr) to make the width fit the display ?

(It is not a monitor issue, I tried to shrink the wide size but the mirroring effect still remains at that resolution)

EDIT: WAIT !! O_O After a reboot it works correctly again...

I really hate those automatic configurations at every boot... some day I will really have to create an xorg.conf... at least I will be sure of the right configuration...
 
Last edited by a moderator:
BEWARE !! With the last "ntfs-3g" security update the hell raised in my PC...

It took a while to find out that was the cause... because that simple update did this:

1) Scary errors at the boot (with kernel 3.16) regarding writing outside the disk limit, followed by a cool KERNEL CRASH

2) I was able to boot with kernel 3.2 but found out that many services weren't able to start (including the bluetooth one)... also, I was unable to mount all my external HDs and even the one with Windows...

Downgrading the "ntfs-3g" to the previous version SOLVED EVERYTHING:

Code:
sudo apt-get install ntfs-3g=1:2014.2.15AR.2-1
 
Last edited by a moderator:
Hmm... I use Debian Jessie on several ARM and x86 devices. Use the latest Fedora on PPC64, ARM and X86 platforms and don't have half of the issues you seem to find.
 
Last edited by a moderator:
I've even installed arch on two machines and had plenty of bugs and troubles to work around, but it's felt a lot more straightforward than the troubles you've had, Powergod.  I do run only intel graphics though, which at least means I can usually see what I'm doing.
 
Maybe just because my machine is old

EDIT: ...and here again the 640x480 issue... now it doesn't go away even after some reboot... :(
 
Last edited by a moderator:
Ok, I will NEVER update Debian on this PC AGAIN !!

I just have to find out what is the cause of the kernel panic at the boot this time... :(

ntfs-3g was locked with the working version, it was not updated, but I have the same behaviour...

EDIT:

I suppose the problem is with these:

Code:
 libudev1:i386 (215-17, 215-17+deb8u1)
 udev:i386 (215-17, 215-17+deb8u1)
 libgudev-1.0-0:i386 (215-17, 215-17+deb8u1)
but the previous version is not on my system anymore O_O and the trick I used before does not work... :(
 
Last edited by a moderator:
Well, I can't access anymore 3 hard drives from linux... and I have everything on those... an OS that can't access data is not so useful...

I tried to update ntfs-3g just to see if sending it back to the old version should have corrected problems with other packages...

BUT, after the update it's old version is gone !! I can't downgrade it anymore O_O

How can I get back the old versions if them are not anymore listed with apt-cache ?

EDIT:

I added all the possible Debian repository in the world, oldtable/stable/wheezy/jessie/testing/etc...

But the not patched version of ntfs-3g is not anymore there !! There's only the 2012 version and it's too old, I had the 2014...

Where the hell are gone all the versions previous to the last upgrade ?! o_O

EDIT:

Because I was out of options I downgraded ntfs-3g to the 2012 version (oldstable)...

after a reboot, not only I was able to boot from the last kernel without crashes, but everything is now working faster than before o_O

Well... seems like a good ending... now I just have to clean a bit my sources.list file...
 
Last edited by a moderator:
I'm not so familiar with the debian way of doing things, but your cache is probably somewhere under /var, and I wouldn't have expected it to have been purged yet.

If you can find the old .deb file anywhere (locally or online) then you can install it using something like 'dpkg -i' from memory.
 
I'm not so familiar with the debian way of doing things, but your cache is probably somewhere under /var, and I wouldn't have expected it to have been purged yet.

If you can find the old .deb file anywhere (locally or online) then you can install it using something like 'dpkg -i' from memory.
As far as I saw from my trials, the list of versions is taken from the repo, if they remove them from the repo and you do an update, you can't find them anymore...

https://packages.debian.org/search?keywords=ntfs-3g

The version "1:2014.2.15AR.2-1" is not anymore there...

EDIT:

WAIT, I actually found something in /var/cache

"./apt/archives/ntfs-3g_1%3a2014.2.15AR.2-1_i386.deb"

EDIT:

Installed, but the KERNEL PANIC started again at the boot, so it must be some connection with ntfs-3g and the others new updated packages...

I'll stick with the 2012 version for now, seems working well anyway
 
Last edited by a moderator:
I don't know when this thing started, I never seen it before, but there's a process that is making the PC totally unresponsive, and writing the hell on the HD...

Seems like related to Synaptic, but I don't use it anyway, so I made the calling script not executable:

Code:
sudo chmod 644 /etc/cron.weekly/apt-xapian-index
 
Back
Top