Occasionally boots into "low graphics mode"

I have the same problem, with a different error message sometimes:
Failed to allocate primary surface: out of memory

I've tested the mainline daily current 2.6.24-999 kernel as suggested in bug 521260 https://bugs.launchpad.net/ubuntu/+source/linux/+bug/521260 , but nvidia-current won't build on it.

As a temporary workaround, I have the nvidia-current driver mostly working and booting without error, but I've had to change the grub gfxpayload to 640x480. X works properly, but my VTs look terrible now (and I spend a lot of time in VT). (Found this in post #4 by mrowth in the forum thread above)

I'm marking this as confirmed.

Also, this only started happening around beta2, either the *-21 or *-20 kernel. Prior to that everything was fine, but I reinstalled hoping to fix the problem (before finding this bug) and haven't been able to revert to older kernels yet.

I am exeriencing following error:

(EE) Apr 18 17:40:33 NVIDIA(0): Failed to allocate primary surface: out of memory

as soon as I have enabled "set gfxpayload=keep" in /etc/grub.d/00_headers, in order to have a better resolution of plymouth splash screen.

Removing that line, I don't have any nvidia error anymore, but my plymouth is horrible.

Thank you for confirming this. Any updates or better fixes to this besides removing a line?

The line to keep gfxpayload isn't in /etc/grub.d/00_header by default,
but VT looks terrible without it. /etc/defaults/grub sets the graphics
mode to 640x480 by default as well.

Also, yesterday I got thrown into low graphics mode even with 640x480
VTs. This hasn't happened before, but proves that setting the VT to
640x480 helps, but dosn't solve, the problem.

according to the ati bug above, this is possibly a kernel issue, and
I'm thinking it might be grub. Should we subscribe them as well? We're
running out of time and this is a showstopper for me.

also, where can i find a deb of kernel 2.6.32-18 or -19? I don't think
they had this problem.

On 4/21/10, Aries K <email address hidden> wrote:
> Thank you for confirming this. Any updates or better fixes to this
> besides removing a line?
>
> --
> Occasionally boots into "low graphics mode"
> https://bugs.launchpad.net/bugs/562565
> You received this bug notification because you are a direct subscriber
> of the bug.
>

--
Sent from my mobile device

Yes I agree this is a showstopper. I am surprised more people are aren't up in arms about the whole situation. Asking for a smooth restart is not asking for too much. I guess we might as well subscribe to the grub bug as well and let this get more exposure. Do you have a link to it? I hope we get the newer .33 kernels so this can be stopped once and for all. :(

I tried booting into the -19 kernel and had the same issue, I'll take out the gfxpayload line and see if it helps

Must be a grub issue, then.

On Fri, Apr 23, 2010 at 12:11 PM, Pierce <email address hidden> wrote:

> I tried booting into the -19 kernel and had the same issue, I'll take
> out the gfxpayload line and see if it helps
>
> --
> Occasionally boots into "low graphics mode"
> https://bugs.launchpad.net/bugs/562565
> You received this bug notification because you are a direct subscriber
> of the bug.
>

Same intermittent problem here with geforce go 7300 in dell laptop. But, I did not have any problem until I add to /etc/default/grub:

GRUB_GFXMODE=800x600
GRUB_GFXPAYLOAD_LINUX=keep

and then running sudo update-grub.

In other words, with GFXMODE disabled, I successfully started X. But, the reason I tried using GFXMODE was because I found it as a recommendation to fix the oversized splash screen on boot.

I guess I could live with the splash screen problem it I had to, but I don't use the VT's very much (actually, only when troubleshooting X problems.

hope there's a work around.

I think I found a solution in https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/551013

Don't use "GRUB_GFXPAYLOAD_LINUX=keep" and instead use v86d module:

aptitude install v86d

edit /etc/default/grub and add after word "splash" in the GRUB_CMDLINE_LINUX_DEFAULT line: video=uvesafb:mode_option=<desired resolution>-32,mtrr=3,scroll=ywrap

My whole line is:
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash video=uvesafb:mode_option=1280x1024-32,mtrr=3,scroll=ywrap"

sudo update-grub

I've rebooted dozens of times with this and no longer have X fail.

You may have to experiment with size .. I tried 1680x1050 and system failed to complete boot. But 1280x1024 is still better than default, and now boot splash is readable again.

hth

Pauls, thanks for the link! Adding that line to just grub didn't work, but the following did (from mabawsa, comment #32 in bug 551013):

>sudo gedit /etc/initramfs-tools
>
>add:
>uvesafb mode_option=1280x800-32 mtrr=3 scroll=ywrap
>
>obviously changing your resolution to what you want then:
>
>sudo update-initramfs -u

And then I followed pauls directions above for good measure. :D

Everything is working well on my nVidia GeForce 6200 OC now, although I think the boot time has increased a bit. That's not a big deal though.

Thanks!

I tried Pauls' method and still get the occasional x-fail. Afterwards, I tried Ryan Waldroop's method and when gedit pops up I get "/etc/initramfs-tools is a directory." so I have nothing to edit. Can anyone please tell me if I did something wrong? Thanks again.

I think Ryan meant /etc/initramfs-tools/modules file. Try that.

regards

Just curious, can you post the default menuentry from your /boot/grub/grub.cfg file. Mine looks like this:

menuentry 'Ubuntu, with Linux 2.6.32-21-generic' --class ubuntu --class gnu-linux --class gnu --class os {
        recordfail
        insmod ext2
        set root='(hd0,7)'
        search --no-floppy --fs-uuid --set 2b87ed9a-1d92-4416-9bf3-6f606c9ab08d
        linux /vmlinuz-2.6.32-21-generic root=/dev/mapper/vgsda8-lvsda8root ro quiet splash video=uvesafb:mode_option=1280x1024-32,mtrr=3,scroll=ywrap
        initrd /initrd.img-2.6.32-21-generic
}

Aries/Pauls, yes, I meant /etc/initramfs-tools/modules

Here's my menuentry from grub.cfg:

menuentry 'Ubuntu, with Linux 2.6.32-21-generic' --class ubuntu --class
gnu-linux --class gnu --class os {
        recordfail
        insmod ext2
        set root='(hd1,1)'
        search --no-floppy --fs-uuid --set
5a8ae36a-755c-458c-b56e-d81baabe8295
        linux /boot/vmlinuz-2.6.32-21-generic
root=UUID=5a8ae36a-755c-458c-b56e-d81baabe8295 ro quiet splash
video=uvesafb:mode_option=1024x768-32,mtrr=3,scroll=ywrap
        initrd /boot/initrd.img-2.6.32-21-generic
}

I have the same bug as you guys descripted above.

Patrick: do either of the above posted workarounds work for you?

On 4/28/10, Patrick Vijgeboom <email address hidden> wrote:
> I have the same bug as you guys descripted above.
>
> --
> Occasionally boots into "low graphics mode"
> https://bugs.launchpad.net/bugs/562565
> You received this bug notification because you are a direct subscriber
> of the bug.
>

--
Sent from my mobile device

Here is my default menuentry:

menuentry 'Ubuntu, with Linux 2.6.32-21-generic' --class ubuntu --class gnu-linux --class gnu --class os {
 recordfail
 insmod ext2
 set root='(hd0,1)'
 search --no-floppy --fs-uuid --set 7091c65c-bcd6-4fc7-b3ba-437021402efe
 linux /boot/vmlinuz-2.6.32-21-generic root=UUID=7091c65c-bcd6-4fc7-b3ba-437021402efe ro quiet splash video=uvesafb:mode_option=1280x1024-32,mtrr=3,scroll=ywrap
 initrd /boot/initrd.img-2.6.32-21-generic
}

Update: Just tried Ryan Waldroop's method of editing the "/etc/initramfs-tools/modules" file and it dramatically helped my problem although it did not get rid of completely. I am now able to reboot more often without getting the x-crash. I rebooted 4 or 5 times in a row and finally got x to crash versus before I was lucky to get 2 good boots in a row. I used 1280x1024 instead of 1280x800, do you think changing the resolution to 1280x800 will help even more? Glad to have made some progress though. :)

Aries: set the resolution to 1024x768 for a standard 4:3 square monitor. Or
1280x800 for a widescreen.

On Apr 29, 2010 5:45 PM, "Aries K" <email address hidden> wrote:

Update: Just tried Ryan Waldroop's method of editing the "/etc
/initramfs-tools/modules" file and it dramatically helped my problem
although it did not get rid of completely. I am now able to reboot more
often without getting the x-crash. I rebooted 4 or 5 times in a row and
finally got x to crash versus before I was lucky to get 2 good boots in
a row. I used 1280x1024 instead of 1280x800, do you think changing the
resolution to 1280x800 will help even more? Glad to have made some
progress though. :)

-- Occasionally boots into "low graphics mode"
https://bugs.launchpad.net/bugs/562565 You receive...

Edit: After numerous restarts I noticed the x-crashes still continue for me. I put the resolution at 1280x800 since I have a wide-screen. Earlier when I mentioned having 4 successful restarts I guess it was just luck since x is crashing like before so disregard what I said about it "dramatically fixing my problem". I tried both workarounds mentioned.

NVIDIA GeForce 8600GM 512 MB

I tried Ryan Waldroop's and Paul's methods and neither worked for me. The error still remains the same. Further I'd like to report that occasionally my entire system freezes, and the only option I am left with to recover it (after having tried killing off all programs individually from a console tty it still does not recover) is to restart the gdm service. I'm wondering if this is in any way connected, not sure. This is probably an exclusive bug so I'm not going to post the relevant portions of my syslog here unless needed.

As to this bug, it still remains, neither of the fixes worked for me, I tried it with a number of resolutions. I have now reverted back to NVIDIA Driver version 173.14.22, the latest being 195.something

beadrifle, no freezes here after a few weeks on lucid. But, this is my "productivity" machine, so I chose conservative options like 32-bit and ext3 and gnome.

I'm running an x64 build of lucid here on ext4; using gnome too.
The freezes persist even after downgrading to 173.14.22, I am yet to compare the syslogs after downgrading the driver though.

The new kernel update is out (2.6.32-22-generic) but the problem still persists. Same issues when I upgrade to version 195 of the NVIDIA driver. even 173.14.22 which seemed fine does throw up issues and is tremendously slow. The only version that seems to work fine is the recommended version (170).
Bump.

Hi there!

I had the same problem (or I think I don't have any more).
I used to have the CPU frequency scaling monitor applet, and since I removed it, I can boot normally. Maybe there is some sort of connection here.
Do you people have this applet or similar ones?

Hi again.
Looks like it was just a coincidence because I still have this problem. :(

I was really hoping the new kernel would have fixed things. Very 170 that
I'm using right now isn't all that smooth, there're still white streaks that
I can see now and then just below my window's bar at the top when I drag the
window (high desktop effects enabled).

On Sat, May 8, 2010 at 3:48 AM, Ricardo Rodrigues <email address hidden>wrote:

> Hi again.
> Looks like it was just a coincidence because I still have this problem. :(
>
> --
> Occasionally boots into "low graphics mode"
> https://bugs.launchpad.net/bugs/562565
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in Ubuntu: Confirmed
>
> Bug description:
> When I restart sometimes my machine boots into "low graphics mode". Usually
> after a couple of restarts the problem fixes itself and everything is A-ok
> unless I restart (it MAY take several restarts to trigger problem). My specs
> are:
>
> Acer M1640
> -1.8GHZ Intel Dual Core
> -4GB of Crucial ram
> -Nvidia 7050 Onboard Graphics
> -250GB HD (Ubuntu)
> -160GB HD (80GB X2 Raid 0/1) (Windows 7)
>
> The screens read as follows:
>
> "Ubuntu is running in low-graphics mode
>
> The following error was encountered. You may need to update your
> configuration to solve this.
> (EE) Apr 13 16:26:54 NVIDIA (0): Failed to initialize the NVIDIA kernel
> module. Please see the
> (EE) Apr 13 16:26:54 NVIDIA (0): system's kernel log for additional error
> messages and
> (EE) Apr 13 16:26:54 NVIDIA (0) Consult the NVIDIA README for details.
> (EE) NVIDIA (0): ***Aborting***
> (EE) Screen(s) found, but none have a visible configuration."
>
> The screen that follows the aforementioned one reads:
>
> "What would you like to do?
>
> -Run Ubuntu in low-graphics mode just for one session
> -Reconfigure graphics
> -Troubleshoot the error
> -Exit to console login
> -Restart X"
>
> Other users have had this problem as well:
> http://ubuntuforums.org/showthread.php?t=1451351
>
> This maybe a potential show-stopper for some NVIDIA users as restarting
> frequently can be frustrating.
>
> To unsubscribe from this bug, go to:
> https://bugs.launchpad.net/ubuntu/+bug/562565/+subscribe
>

At this point probably the only thing that will alleviate this problem is the 2.6.33-xx kernel. Unfortunately we won't be seeing this in Lucid for quite a while :(.

Two new updates came out the day before; on nvidia-current and nvidia-modaliases packages. If you have been using the propriety drivers that come along with ubuntu (version 170) then after these updates your driver is brought up to date with the latest (version 195) and there is no issue of booting into low graphics. I can confirm this.

Booting however takes longer than before, and the boot ubuntu screen still shows up in low resolution by default, but this I believe can be fixed. I can confirm this bug to be fixed (at least for now) but with some side effects on the bootup time.

I have same issue (installed all updates). Without GRUB_GFXPAYLOAD_LINUX=keep it doesn't crash but then tty's are not available (bug #447765). I need to issue `sudo restart gdm` command after booting into low graphics mode and now nvidia is loaded perfectly. This is really annoying, I am having these issues (bug #447765) since kermic but no fix yet.

@ beadrifle

No those two updates do not fix the problem for me. I still occasionally boot into low graphics mode.

https://bugs.launchpad.net/ubuntu/+bug/562565/comments/10 and https://bugs.launchpad.net/ubuntu/+bug/562565/comments/11 (paul and Ryan Waldroop's posts on this thread) have improved this bug for me quite a bit. Now I only have to reboot once or twice instead of the 10+ times that I was having to in the past couple of weeks. Although that changes this from an absolute show-stopper to an annoyance, I am still interested to see if there is a solution short of waiting for 10.10 and the new kernel. Are there people who have tried kernel 2.6.33? Was it to solve this issue?

Have you tried the latest of the nvidia-current and nvidia-modaliases packages? After I got the update on these the default nvidia drivers that come along with the lucid bundle (170) upgraded to the latest (195) and since I have not faced this issue.

Good day,
I'm trying to install the nVidia Linux driver, version 195.36.31 and 195.36.24, on Ubuntu 10.04 i32 (Lucid Lynx).
I've tested the installation against the kernel version that shipped with the distro (2.6.32-21) and 2.6.32-22.

On both kernel versions the nvidia.ko kernel module loads, but when I attempt to launch the X server the process fails. The following error is logged to /var/log/messages:

   NVRM: RmInitAdapter failed! (0x27:0x28:1094)
   NVRM: rm_init_adapter(0) failed

I've made sure that I've deinstalled the nouveau package and blacklisted the module. Also, some threads suggested that "vmalloc=256M" should be added to the GRUB default cmd line. I've confirmed that the size is increased when I 'cat /proc/meminfo', but that doesn't help either:
   VmallocTotal: 524288 kB
   VmallocUsed: 34920 kB
   VmallocChunk: 480892 kB

It does seem asif the PCI ID is correctly detected, if I compare the bus id in xorg.conf with the value from lspci:
01:00.0 VGA compatible controller: nVidia Corporation GT218 [GeForce 210]
(rev a2)

This seems to be a problem specific to the GeForce 210 cards

Temporary Solution: Reverted back to nVidia driver 195.36.15

I've attached the nvidia bug report. Maye there is something in there that can shed some light on the issue of why 195.36.15 works and 195.36.31 doesn't work for GeForce GT218 (GeForce 210)

I use the nvidia-current package 195.36.24-0ubuntu1~10.04 and nvidia-current-modaliases 195.36.24-0ubuntu1~10.04. My card is a Geforce 210. I haven't had any Xserver crashes recently but I can't say since when this seems to have been solved.

I agree, I have not had an Xserver crash for a good while. It is safe to assume it is fixed.

This bug is still affects me, occasionally giving the "low graphics mode" error. Running kernel 2.6.32-24 (current) and the recommended nvidia driver 195.36.24 (also current). It's from my macbookpro w/ Nvidia 9400. Seeing that I use ubuntu quite a bit here at school, this bug is definitely a show stopper. Any ideas?

Can you run "apport-collect 562565"?

We're closing this bug since it is has been some time with no response from the original reporter. However, if the issue still exists please feel free to reopen with the requested information. Also, if you could, please test against the latest development version of Ubuntu, since this confirms the bug is one we may be able to pass upstream for help.