nvidia-drivers-304.x drivers aren't updating the kernel mod?

ExecutorElassus

So, after a lot of other problems with a dying video card, I'm back to being able to use the 3xx series of nvidia-drivers. Thing is, I don't seem able to use anything beyond the 302.17-r1 driver: both of the recent 304 drivers install fine, but I can't start X afterwards, getting an API mismatch error (ie, that the X driver component is the new version, but the running kernel module is the old one).

The usual solution to this is to unload the kernel module, then reload it, yes? That doesn't seem to work. 'eselect kernel list' shows the currently-running kernel is the active one for symlinking as well.

Is there a step I'm missing here?

Cheers,

EE

NeddySeagoon · Posted: Sun Sep 02, 2012 11:49 am Post subject:

ExecutorElassus,

Everything you say is correct, yet it contains a contracdiction somewhere.

Please post the output of

ExecutorElassus · Posted: Sun Sep 02, 2012 11:59 am Post subject:

Hi Neddy!

Nice to hear from you again!

Here are the two commands:

NeddySeagoon · Posted: Sun Sep 02, 2012 12:05 pm Post subject:

ExecutorElassus,

uname -a tells about the running kernel - its 3.5.0-gentoo, built on Tue Jul 31 15:31:38 CEST 2012.
readlink says you are building external kernel modules for 3.5.3-gentoo

The root cause of the problem is that you are not running the kernel you think you are. There is a lot of that about.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

ExecutorElassus · Posted: Sun Sep 02, 2012 1:54 pm Post subject:

Ha ha!

so, it seems that when I was installing my new video card, I knocked the power cable for /dev/sdc* out, and mdadm dropped it from the arrays. Then, sda1 somehow got booted into its own array, so /boot -- being based on one of the mirrored partitions -- was being updated, but grub itself was loading the kernels from one of the other (non-updated) partitions.

So, now the sdc* partitions are back into their respective arrays (yay for six-hour re-sync operations!), and grub is booting the correct kernel.

I really wish mdadm would give more urgent warnings when stuff like this happens. A three-device array getting reduced to a single device should be considered a nontrivial problem.

Anyway, thanks for pointing out what the problem was. It seems to be fixed now.

Cheers,

EE

NeddySeagoon · Posted: Sun Sep 02, 2012 2:05 pm Post subject:

ExecutorElassus,

mdadm can email you about things dropping out of the array.

You can use an array thats in the process of being synced too. mdadm will get out the way and slow the sync.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

ExecutorElassus · Posted: Sun Sep 02, 2012 3:21 pm Post subject:

Hey Neddy,

yes, as you might remember from our last exchange, I have a lot of experience with re-syncing RAID arrays.

I'm trying to set up mdadm to send me messages, but I get the following error:

Jaglover · Posted: Sun Sep 02, 2012 9:05 pm Post subject:

https://forums.gentoo.org/viewtopic-t-412468.html

I personally prefer a local mail server that runs in my router.
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!

ExecutorElassus · Posted: Sun Sep 02, 2012 9:36 pm Post subject:

Fixed it. Had to enable SMARTTLS. Thanks!

I think everything I was having problems doing with this thread is solved. Thanks for the help, y'all.

<3

EE