Kernel panics
- Anmelden oder Registrieren um Kommentare zu schreiben
The following happened on a Gateway MX3410 notepad running Trisquel 6.0 Live from a bootable usb memory. The built-in hardrive with a full and updated Trisquel 6.0 installation was now intentionally excluded from BIOS. The issue continues to happen. At present an external replacement hardrive is being used with no improvement:
Here is some technical information from my current live session:
trisquel@trisquel:~/Documents$ lsblk -f
NAME FSTYPE LABEL MOUNTPOINT
loop0 /rofs
zram0 [SWAP]
sda
├─sda1 /cdrom
└─sda2 /media/2
sdb
├─sdb1
├─sdb5
├─sdb6 [SWAP]
└─sdb7 /media/7
trisquel@trisquel:~/Documents$ sudo blkid
/dev/loop0: TYPE="squashfs"
/dev/zram0: UUID="9dbf9df4-f0a8-40d5-a0c5-21de61600752" TYPE="swap"
/dev/sda1: UUID="0AA2-CA25" TYPE="vfat"
/dev/sda2: UUID="a22cb343-4768-486a-b747-52ea110013a8" TYPE="ext2"
/dev/sdb5: LABEL="STORE-IT" UUID="3BF5-4A00" TYPE="vfat"
/dev/sdb6: TYPE="swap"
/dev/sdb7: UUID="3dbfb20f-cecc-4ea4-bef6-f414e0e90452" TYPE="ext4"
trisquel@trisquel:~/Documents$ uptime
23:13:58 up 11:56, 8 users, load average: 0.17, 0.24, 0.22
trisquel@trisquel:~/Documents$ free -m
total used free shared buffers cached
Mem: 368 363 4 0 12 69
-/+ buffers/cache: 281 87
Swap: 983 192 790
Here is what was on the screen before being forced to shutdown earlier today:
1424160805 /media flashingCapsLock-GatewayPc-20150217-09h59.t The following information is manually typed in after resuming from suspend.
The first eight lines end with [ OK ] on the extreme right (minus 1 space), immediately followed by a blank line and a ninth [ OK ]
The second blank line ends with [ on the extreme right followed by 21 lines with values starting fromm 7797.409769 to .410061
-------------------- top of screen ----------------------------------- ...
* Stopping anac(h)ronistic cron
* Stopping cold plug devices
* Stopping log initial device creation
* Starting enable remaining boot-time encrypted block devices
* Starting configure network device security
* Starting configure virtual network devices
* Stopping configure virtual network devices
* Stopping enable remaining boot-time encrypted block devices
[
7797.409769] Kernel panic - not syncing: Attempted to kill init!
[ 7797.409803] Pid: 1, comm: init Tainted: G G 3.2.0-38-generic-pae #0trisquel1
[ 7797.409816] Call Trace:
[ 7797.409838] [] ? printk+0x2d/0x2f
[ 7797.409850] [] panic+0x5c/0x161
[ 7797.409868] [] forget_original_parent+0x1e4/0x1f0
[ 7797.409883] [] ? _raw_spin_lock_irqsave+0x2d/0x40
[ 7797.409896] [] exit_notify+0x13/0x100
[ 7797.409907] [] do_exit+0x1ae/0x3c0
[ 7797.409918] [] do_group_exit+0x38/0xa0
[ 7797.409934] [] get_signal_to_deliver+0x1b6/0x3f0
[ 7797.409947] [] ? vmalloc_fault+0x195/0x195
[ 7797.409961] [] do_signal+0x0x3f/0xd0
[ 7797.409975] [] ? pipe_write+0x25b/0x4a0
[ 7797.409985] [] ? do_page_fault+0x418/0x490
[ 7797.409997] [] ? vfs_write+0xed/0x160
[ 7797.410007] [] ? wait_on_retry_sync_kiocb+0x50/0x50
[ 7797.410019] [] ? _raw_spin_lock+0xd/0x10
[ 7797.410030] [] do_notify_resume+0x75/0x90
[ 7797.410041] [] work_notifysig+0x13/0x1b
[ 7797.410061] panic occured, switching back to text console
To me it seems that it is just lacking in RAM.
Trisquel's requirements can be found here (http://trisquel.info/en/wiki/system-requirements) , but you will have to confirm that you are using the default ISO (uses GNOME).
The processor seems to be 64-bit (http://www.cpu-world.com/CPUs/K8/AMD-Turion%2064%20X2%20Mobile%20technology%20TL-50%20-%20TMDTL50HAX4CT%20%28TMDTL50CTWOF%29.html). Confirm that the link corresponds to your model.
There seem to be other versions of that laptop with 1GB RAM, so a RAM upgrade is very recommended.
Lastly, you should switch to using the newer 7.0 Mini ISO that comes with LXDE (desktop environment), which runs with the Openbox (window manager). If you are planning on keeping a GUI you probably want to switch to an even lighter window manager.
if ram is a problem trisquel mini will run on 128mb of ram
here are the minimum system requirements:
https://trisquel.info/en/wiki/system-requirements
Just finished downloading and trying trisquel mini. Cannot find option to disable mouse clicks with touchpad. After disabling automount, was getting error pop ups on the desktop for every media change but /media/trisquel stayed clean after umounting those universal id's initially placed there on startup. Memory usage seems unchanged between mini and standard edition where I can do screenshots and document stuff with gedit. After resuming from suspend I was getting i/o errors for system commands including man and sudo fdisk -l. Other commands such as lsblk continued to work and showed that /dev/sdb's had, again migrated to /dev/sdc's. This never ever happened on another pc running trisquel live for an uptime already exceeding one year with trisquel 6.0 live! I suspect "Strong ROM" intialization has been corrupted on this Nvidia drive controller that somehow is tightly knit with BIOS. Gateway seems tightlipped about this annoyance. Subsequent attempts to start trisquel lite from usb repeatedly end up with a black screen now with a flashing cursor in the upper left hand corner. Replugging trisquel 6.0 live got me here to follow this up. Also, option to test memory is blocked from the live boot menu.
Thank-you for your assistance. When suspending on live version, the system is able to recover but mount points for /dev/sdb.. are invalid. Checking with sudo fdisk-l they have migrated to /dev/sdc.. remounting restores operation until next suspend, then I have to remount the same to /dev/sdd.. Is this correct?
Also the previously woking memory test feature on the live usb is disabled and a window entitled in grey: Boot loader; and contents in the white portion: "/isolinux/mt86plus: file not found" pops up.
I never had this when booting from the same usb before on this machine. What could have caused this? The md5 sums haven't changed on this or a new USB made from the running live version.
/dev/sdX (where X is a letter) is not a mount point (that must be a directory). It is a device. It should contain some partition(s) /dev/sdXY (where Y is a number). Such a partition can be mounted wherever you want. For instance, to mount the partition 1 of the disk b on /mnt:
$ sudo mount /dev/sdb1 /mnt
That said, suspension should not create such problems. Are we talking about partitions mounted at init? If so, how are they identified in the first column of /etc/fstab (by UUID, by label or the device node)?
Memtest86+ is not on Trisquel ISOs. That means you cannot test your memory with the Trisquel live system (well, actually, you can by installing "memtester" that does not require a reboot). You can install the "memtest86+" package on Trisquel though.
One thing you could try to mitigate that is edit the suspend scripts to unmount and the resume scripts to mount the devices.
I remember seeing that error message once and editing some text file on the usb stick changing isolinux to syslinux in the path. I believe this might be related to what tool you used to make the usb.
Thank-you for your time and assistance.
The usb I made was from the utility included with the Trisquel Live 6.0 System settings menu, using an .iso file and opting out of persistent storage for it.
I have not yet looked for any suspend scripts, in the ram-resident file system. I will try that next!
Thank-you.
I stand corrected in reference to mount points. When running live, the device list changes after each resume from suspend. I specified "do-nothing" under system power options when closing lid, but system goes into suspend nevertheless. Device UUID's remain unchanged.
With the built-in hardrive physically removed and BIOS device list specifying "none" for it, am now running a Winchster usb device (/dev/sdb) with an ext4 partition number at (/dev/sdb7) that is manually mounted to a manually specified subdirectory made with a simplified name under /media. I/O errors occur and the automatically assigned swap partition to /dev/sdb6 becomes /dev/sdc6! I will clearly run out of device letters from each resume from suspend this way. Another, new, non-internet machine running Trisquel 6.0 live has an uptime exceeding 368 days where my usb drives remain unchanged and active during and after many suspends.
Regarding Memtest 86+ it worked repeatedly before first installing Trisquel 6.0 on this Gateway MX4310 Windows-only home machine, after shrinking a defragmented ntfs factory installed partition on addresses following the FAT32 recovery partition. After nearly a hundred Kernel panics, and occasional lockouts preventing me from even getting to grub, I could coax the Windows system to life just to see if it still worked and existed. First with startup errors, later with a few blue screens and now with a black text screen complaining about a missing or corrupted Windows file! Before blaming the hard drive, am seeking live performance consistency, without kernel panics, that still have a tendency to occur even when leaving everything switched on overnight...
It looks like you have an hardware issue. I hope you frequently backup your data! Have you recently "auto-tested" your drive? You can do that from the "Disk" utility in "System settings" (use the menu button in the upper-right corner of the windows).
I would try to properly define permanent mount points for your external devices ans see if that solves your problem: https://help.ubuntu.com/community/RenameUSBDrive
Thank-you again for your valuable time helping me:
Indeed I have a hardware issue: I would like to bypass zram0 [SWAP] which makes up 184.1M of total Swap memory of 983M; where the Disk utility 'sees' only the /dev/sdb6 [SWAP] |-sdb6 of 799M which was wiped using DoD 5220.28-STD (7 passes) and reformatted as a Linux swap.
The vast majority of Kernel panics would occur when I allocated the BIOS default of 64M frame buffer for Nvidia. What I suspect is that the unallocated portion is used as swap space.
To veriy this, I went back from the maximum allowable 128M to 64M frame buffer allocation in the BIOS and saved settings and exited and surely enough Kernel panics would occur with both usb bootable devices, with no drives in any slots, the Minitool partitioning tool alone or Trisquel alone respectively. Cutting the power and trying once again got me into the partitioning tool, and then I plugged the removable drive with the 799M swap partition, refreshed the device list and proceeded to wipe it.
With the Trisuel usb, I got kernel panic after selecting try Trisquel. Cutting power and trying that once again got me into Trisquel - momentarily a high resolution black screen appeared with one line message: ATA failed - invalid argument, and I was back here!
Now zram0 has 215.8M capacity and total swap space, adding the 799MB, indeed is 1014MB. Is there any way using the "Disk utility" to disable and test the drive which is embedded in the video controller? I have about 4MB memory available after booting into the Desktop when the swapper is still vacant. I would like to try reassigning swap space to the sdb6 799MB device only. How?
I do not understand much of what you wrote. I was only suggesting you to test the disk. The whole disk. If the disk is failing, you need to change it. There is no fixing.
The swap is some disk space (a partition in your case but swap files exist too) used as (a very slow) main memory when the RAM is full. You can disable the use of all the swap with 'sudo swapoff -a' (valid until the next reboot). You can also control the relative weight given to swapping out runtime memory: https://en.wikipedia.org/wiki/Swappiness
From a Live media (such as Trisquel's ISO, which includes GParted), you can create/delete/enlarge/shrink partitions (including a swap partition) but be careful: /etc/fstab (that of the installed system, not that of the live system) must be changed accordingly.
As far as I understand, the framebuffer is in the RAM (nothing to do with the disk, even if the system swaps). When you reduce it from the BIOS, you increase in parallel the RAM the system can use and decrease the use of the swap. As a consequence, if the disk is the culprit, the IO errors may be delayed thanks to that. Anyway, you do not want to use a failing disk!
That said, because the problem occurs as well with a live system, it might more probably be a problem with the RAM... although I think GNU/Linux live systems can use the swap partition on another disk in the system (I am not sure but you could run the live system after disconnecting the internal drive just to see if the problem disappears).
Anyway, 1 GB of RAM is too little to run the normal edition of Trisquel (that is why it swaps). You want Trisquel Mini.
Trisquel gnome flashback runs perfectly on my lappy 1 gb ram.
Sure if I open 10 apps or even 5 heavy ones.. swap swap swaaaap
Magic Banana: Thank-you for getting me this far. The live system boots without any hard drives present leaving me with 4 megabytes free RAM. A solid state drive, not visible in gparted, was found using the disk utility. All of the space appeared as grey and unallocated (approx. 200MB) and capacity corresponding to the reported swap space. Also, what was taken away for frame buffer allocation in BIOS settings became available by a proportionately increased amount on the solid state drive and larger reported swap space.
With an open terminal I did free -m and saw swap space was still unused before trying to benchmark the solid state drive using the read/write method. It could not be started because it was already being used by now. I successfully benchmarked it read-only. Retrying read/write method thereafter permitted me to start benchmarking and each time errors appeared: twice at one address then at at a higher one, ending with a statement that the benchmarking exceeds disk capacity. Repeating free -m finally showed swap usage exceed 0. By now I had lost usage of man, sudo and other commands. Bingo!
I doubt you are looking at the proper row of the output table returned by 'free' (1 GB used after the boot is a lot). You must look at the row "-/+ buffers/cache".
It is very weird that "Disks" sees a SSD that GParted does not! Are you sure it is not the live USB you booted?
175mb used after booting in trisquel for me: powaaaaaaaaaa :)
I believe I have isolated the cause of all the weird problems.
Once first up with Trisquel 6.0 Live, I did the following 3 commands:
$ free -m
Mem: 431 425 5
buffers 176 254
Swap: 1014 0 1014
$ sudo swapoff /dev/zram0
$ free -m
Mem: 431 425 5
buffers 178 252
Swap 798 0 798
-------------------------------------------------
With Disk utility, I have the following listing:
Storage Devices
Local Storage
Pata Host Adapter
Peripheral Devices
701 MB File (filesystem.squashfs)
Generic Flash Disk (unused ext2 partition)
40 GB Hard Disk (with fat32, 800MB swap, and Ext4)
193 MB Solid State Disk
I was now able to access and reformat what was still defined as swap on the solid state drive (/dev/zram0) to ext2 and I also encrypted it with a long key, to ensure nothing from it could be accessed by any process. I didn't even bother benchmarking it any more. While I am up it should hopefully stay quarantined.
-------------------------------------------------
With gparted, the only listed devices are the bootable usb flash and usb winchester:
/dev/sda
/dev/sdb
--------
I thank you all for your interest and helpful advice.
Martins
trisquel@trisquel:~/Documents$ uptime
12:12:23 up 16:41, 8 users, load average: 0.00, 0.02, 0.07
trisquel@trisquel:~/Documents$ free -m
total used free shared buffers cached
Mem: 368 353 14 0 16 112
-/+ buffers/cache: 224 144
Swap: 798 138 660
Peace at last!
- Anmelden oder Registrieren um Kommentare zu schreiben