LMDE BREAKAGES - tracking TESTING

[zerozero]

not a problem friend
now you should remove or comment out the repo again, otherwise you gonna get nasty surprises next time you upgrade

[GeneC]

not a problem friend
now you should remove or comment out the repo again, otherwise you gonna get nasty surprises next time you upgrade

Done!

[phil]

I am not sure if this is the place to post my problem, but here goes.

One of my computers, an Intel Atom 610 that has been running LMDE main with Xfce for more than a year recently started having a problem booting after a big batch of updates including grub. The error is from grub "grub_devmod64_full" and it puts me into grub recovery. I can boot with my grub boot disk and the problem seems to be solved after a reinstall of grub, but after several restarts the problem appears again. A search of Google indicates that it is a Debian error #630226 but I can find no solution to fix the error. I will probably be reinstalling LMDE Xfce in the near future but I don't want this to happen with the reinstall.

[hopimet]

Similar problem, not sure how to proceed.

Same issue for me. I have decided to hold the update of these packages and wait.

[mockturtl]

This morning, I wanted to update some packages (see screenshot below) but I have been warned by apt-listbugs of a potential serious bug

Cool find, I didn't know about this package.

Warning, though: I think its prompts ("This serious bug exists: do you want to proceed?") may cause Software Manager to hang indefinitely, awaiting input. (That was my experience installing audacity yesterday, and my best guess at the cause.)

[yaye]

I was able to fix this by temporarily changing from testing to unstable in sources.list, installing nvidia-glx and all dependencies, and then reverting back from unstable to testing. Everything seems to be working great. Did a clean and update and everything still shows up to date and I'm not asked to remove anything that was installed from unstable.

Here's how I did it:

sudo vim /etc/apt/source.list (sub vim with the editor of your choice)
find this line: deb http://ftp.debian.org/debian testing main contrib non-free
change it to: deb http://ftp.debian.org/debian unstable main contrib non-free
save and quit.

sudo apt-get clean
sudo apt-get update
sudo apt-get install nvidia-glx

sudo vim /etc/apt/source.list (sub vim with the editor of your choice)
find this line: deb http://ftp.debian.org/debian unstable main contrib non-free
change it to: deb http://ftp.debian.org/debian testing main contrib non-free
save and quit

sudo apt-get clean
sudo apt-get update
sudo reboot

I don't know if I'm the only one who had problems after following these instructions, but on reboot, Plymouth ran in high resolution as before and then I was dropped to high resolution text mode login. I typed root and root's password, then startx. The xserver gave an error message and I was again returned to the command prompt. This happened late last night and I decided to just shutdown the system and get some sleep.

Today I remembered that my computer has an older nVidia card (GeForce FX 5900) and it requires the 173xx legacy nvidia driver. I used apt-get to remove the nvidia-glx package . I was told that several other packages were installed and they could be auto removed. I individually removed all the other packages that contained "nvidia" in their name (high resolution text mode really came in handy).

I then tried to install the nvidia-glx-legacy-173xx package but it complained that libglx-nvidia-alternatives and libgl1-nvidia-legacy-173xx-glx were required. I installed them in that order. The libgl1-nvidia-legacy-173xx-glx package pulled in some other packages and compiled the nvidia kernel module. When the command prompt returned, I rebooted. Plymouth ran, the xserver started and I was back at the GNOME login screen.

It's probably best to check http://www.nvidia.com/object/IO_32667.html to see if your video card requires a legacy nvidia driver and apt-get install that driver (instead of nvidia-glx) after altering the sources.list .

[MadOCer]

Not really a breakage but perhaps worth noticing:

Wine seems to be gone from the testing repos (as of today???).
I just tried to get some win programs running and experimented with some different versions when I realized that the version formerly available from the testing debian repo is completely gone. Just temporarily pointed the repos to unstable to get it installed. Although it's an easy fix, this is quite puzzling.

Update:
I start to understand why it's stricly "unstable" ATM. Wine just f***ed me over big time. I had a movie encode running for over 12 hours already. Then wine caused a crash, I got logged off and all my running processes got killed. Way to go! How I love it.

[doktordave]

I've noticed workarounds for Nvidia, and was wondering if anyone has come up with one for fglrx/Catalyst drivers yet.

One other quick question - does the breakage occur with all kernels, or does it only affect the 2.6.39 variety?

[kmb42vt]

I've noticed workarounds for Nvidia, and was wondering if anyone has come up with one for fglrx/Catalyst drivers yet.

One other quick question - does the breakage occur with all kernels, or does it only affect the 2.6.39 variety?

Yes, breakage can occur with any kernel. The thing that also needs to be kept in mind that some breakage can be due more to certain hardware configurations as well (and sometimes who manufacturers said hardware as well) than to the kernel itself. As far as the 2.6.39 kernel I can only give an example of the Ubuntu based version that I have previous experience in using. The 2.6.39-0 had a major installation bug in it, mainly that it couldn't be installed due to a bug in the "headers" package. The 2.6.39-1 kernel solved that particular problem and dropped resource usage by 100 MBs at login as compared to the 2.6.38 kernel (which had it's own share of problems) but exhibited many problems in GUI rendering and video streaming once installed that were just random enough to make them very hard to pin down (that's with a supported Nvidia GT-430 card using the up to date drivers). The 2.6.39-2 kernel was quickly released which solved all the aforementioned problems and has been running well so far in my case. I also haven't had any problems so far using the 2.6.39-2 kernel (Update package 1) currently installed in LMDE.

The above is just an example based on just experience of course, just to give you an idea about breakages related to particular kernel.

[doktordave]

Thanks kmb42vt, that's the type of info I was looking for regarding the kernels.

[grizzler]

The 2.6.39-2 kernel was quickly released which solved all the aforementioned problems and has been running well so far in my case. I also haven't had any problems so far using the 2.6.39-2 kernel (Update package 1) currently installed in LMDE.

Try to 'safely remove' a USB external harddisk with more than one partition. Be prepared for a kernel panic.

[kmb42vt]

The 2.6.39-2 kernel was quickly released which solved all the aforementioned problems and has been running well so far in my case. I also haven't had any problems so far using the 2.6.39-2 kernel (Update package 1) currently installed in LMDE.

Try to 'safely remove' a USB external harddisk with more than one partition. Be prepared for a kernel panic.

Wow! Now there's a good one. I've been able to remove my thumb drive using "Safely remove drive" but I haven't had my external hard drive hooked up for the past month or so. That being said, I did a few experiments and I can successfully unmount my external hard drive without a complete lock up of LMDE by opening Nautilus and simply clicking on the "unmount" arrow to the right of my listed external hard drive in the left sidebar. The drive unmounts and the system is still stable. Strangely enough, if I right-click on the now unmounted external hard drive it still has "Safely remove drive" listed in the context menu and it shouldn't be since the drive is not mounted.

Is there a particular log entry that actually shows the kernel panic? My LMDE simply locks up and I have to force a "power-off-reset". Once booted back in I can't find anything about the kernel panic in the "Log file viewer". One thing I have noticed during the boot up sequence using the 2.6.39-2 kernel (once I updated LMDE using the new update system that is) is the following udev error (listed in "syslog":

Code: Select all

udevd[3606]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe /sys/devices/pci0000:00/0000:00:1d.7/usb2/2-5 2 3': No such file or directory

Not sure if that has to do with the problem we're seeing but I thought I'd post it anyway. This needs to be reported if it hasn't been already, at least to the "UPDATE PACK 1 FEEDBACK THREAD":

http://forums.linuxmint.com/viewtopic.p ... 19#p444219

If you have the text of the kernel panic it would be great. Meanwhile a work around is to use the "unmount" arrow the the right your listed external hard drive in the left sidebar of Nautilus as I mentioned above.

[zerozero]

I replicated that: 39.2 kernel amd64 system, usb flash drive (one partition) safely removes works as expected, usb HDD (one partition as well) the same option freezes the system, and i have to hard reboot

[kmb42vt]

I replicated that: 39.2 kernel amd64 system, usb flash drive (one partition) safely removes works as expected, usb HDD (one partition as well) the same option freezes the system, and i have to hard reboot

I'm reposting this to the "UPDATE PACK 1 FEEDBACK THREAD" right now as a matter of fact. If you could post your confirmation there as soon as you see my post show up it would be appreciated. Too bad we can't get a log of the kernel panic but I can't find one so far. System too quickly?

[MadOCer]

Now that USB kernel problems got mentioned: Had witnessed the USB-hdd phenomenon first hand, a few days ago.
Furthermore, using more than one optical drive connected via USB seems to be problematic, too. The one error that is quite understandable: Disconnecting a drive without unmounting it caused numerous kernel oops for me. I worked a lot with the drives the last few weeks, so sometimes I just "accidently" disconnected (disconnected the wrong drive, or forgot to "safely remove" and so on) one of the drives. That's how I got to test that out. Shouldn't really happen.

What's really strange though, is that only copying from both drives at the same time, just today crashed the (internal) USB hub that the drives are connected to. And I highly doubt that the hardware is to blame.
BTW, newest 2.6.39-3 based kernel.
Nearly seems as if the scsi driver handling the io is more than just a bit buggy.
So for anyone else using such a combination of drives, it'll probably be a good idea to use an older, known stable, kernel.

[grizzler]

Is there a particular log entry that actually shows the kernel panic? My LMDE simply locks up and I have to force a "power-off-reset". Once booted back in I can't find anything about the kernel panic in the "Log file viewer".

Nope. This one appears to be too 'violent' to be logged anywhere.

This needs to be reported if it hasn't been already,

http://lists.debian.org/debian-kernel/2 ... 00536.html

at least to the "UPDATE PACK 1 FEEDBACK THREAD":

It isn't really part of that as it predates the Update Pack system. My first encounter with it: http://forums.linuxmint.com/viewtopic.php?f=141&t=75848

[kmb42vt]

It isn't really part of that as it predates the Update Pack system. My first encounter with it: http://forums.linuxmint.com/viewtopic.php?f=141&t=75848

True, but I suggested it since the new update system included the 2.6.39-2 kernel which will cause problems with LMDE both now and in the respin as well and Clem will need a heads up if he doesn't already know about this.

[Gerd50]

Don't update now, if you have nvidia graphic. Driver 275.09.07-4 isn't in the testing repos and sid now has .07-5.

That driver is not working on my machine. Had to restore the former partition.

[zerozero]

i went though the kernel logs and couldn't find a thing about it as well, i found references to the usb-hdd connexion of course, but from there until the reboot it's blank;
about being something that should be reported in the Update Pack thread, honestly don't know, doesn't hurt, and the worst that can happen is the posts being removed

[kmb42vt]

i went though the kernel logs and couldn't find a thing about it as well, i found references to the usb-hdd connexion of course, but from there until the reboot it's blank;
about being something that should be reported in the Update Pack thread, honestly don't know, doesn't hurt, and the worst that can happen is the posts being removed

Yeah, that's about the worst that can happen. Still, it seems to me though that there's a good chance the respin might ship with the exact same problem if it ships with the same kernel and that might leave a bad taste in a (new/previous) users mouth? "Safely" remove their external hard drive and their system freezes hard...doesn't exactly sound like a promotional bell ringer to me.

By the way, I'm running LM 11 with the Ubuntu flavored version of 2.6.39-2 and this particular problem doesn't exist.