My unhappy adventures with Linux installation

[Glyph Reader]

Crossing fingers! 

-Glyph Reader

 

[PowerGod]

During the night it downloaded and installed something like 2 Gb of things, and at the end it reported many incompatible "latex" packets installed together... I never installed it, so I'll have to find info about the other programs that require it...

There's one thing I overlooked while restoring Grub, it have no more a countdown to start Debian...

Also the ssh server does not accept connections anymore... I read that a rule for the root user was changed, but I don't use root to connect... anyway the service seems still up on the same "non standard" port

I'll have another busy evening

 

[levi]

Yeah, sshd can take a little configuration before it opens up. If you've upgraded to an incompatible version, it will have reset the configuration for you.


Edit: Although if it's still listening on the configured port, perhaps the configuration hasn't been reset, just the way it reads it.

 

[PowerGod]

After a reboot ssh is fine, also the timer magically appeared in Grub

Now I have shitty watered colors and a vibrating display... I almost can't read what I'm writing... let's return to the video drivers lectures...

Also I still can't understand if it's faster or slower than before...

 

[PowerGod]

Some differences I'm finding...

- the USB hot-mount put the folders in /media/<user>/<device> instead of /media/<device> ... I'll have to save the changes in many applications...

- the BLUETOOTH WORKS FULLY !!! WITHOUT KERNEL PANIC !!! Sometimes the Blueman interface starts to not respond for a lot of time while changing configurations... but in the end it works !! Now I can finally use my phone as a remote control again

- appeared a "Debian" menu with sub-categories, inside the LXDE menu, that ALSO has subcategories... so I have Applications menu both in "LXDE->Applications" AND in "LXDE->Debian->Applications" with different things inside

- all my icons now are thumbnailed previews of the content... maybe I'll try to find how to disable this, at least for the bigger file types...

- now I have an H.265 codec but I suppose it will be slow on this system

 

[PowerGod]

Looking at dmesg I have seen a lot of errors and warning messages about this "/dev/sdb2", telling that it doesn't have a known filesystem or it is corrupt, etc...

Spoiler

[ 3011.862364] EXT4-fs (sdb2): unable to read superblock
[ 3011.868177] EXT4-fs (sdb2): unable to read superblock
[ 3011.873760] EXT4-fs (sdb2): unable to read superblock
[ 3011.982231] FAT-fs (sdb2): utf8 is not a recommended IO charset for FAT filesystems, filesystem will be case sensitive!
[ 3011.982736] FAT-fs (sdb2): invalid media value (0x0d)
[ 3011.982743] FAT-fs (sdb2): Can't find a valid FAT filesystem
[ 3011.989949] XFS (sdb2): Invalid superblock magic number
[ 3012.003498] FAT-fs (sdb2): utf8 is not a recommended IO charset for FAT filesystems, filesystem will be case sensitive!
[ 3012.003867] FAT-fs (sdb2): invalid media value (0x0d)
[ 3012.003874] FAT-fs (sdb2): Can't find a valid FAT filesystem
[ 3012.046011] MINIX-fs: unable to read superblock
[ 3012.051644] attempt to access beyond end of device
[ 3012.051658] sdb2: rw=16, want=3, limit=2
[ 3012.051665] hfsplus: unable to find HFS+ superblock
[ 3012.058169] qnx4: no qnx4 filesystem (no root dir).
[ 3012.064490] You didn't specify the type of your ufs filesystem

mount -t ufs -o ufstype=sun|sunx86|44bsd|ufs2|5xbsd|old|hp|nextstep|nextstep-cd|openstep ...

>>>WARNING<<< Wrong ufstype may corrupt your filesystem, default is ufstype=old
[ 3012.075251] hfs: can't find a HFS filesystem on dev sdb2
[ 3017.494732] EXT4-fs (sdb2): unable to read superblock
[ 3017.500534] EXT4-fs (sdb2): unable to read superblock
[ 3017.506141] EXT4-fs (sdb2): unable to read superblock
[ 3017.521883] FAT-fs (sdb2): utf8 is not a recommended IO charset for FAT filesystems, filesystem will be case sensitive!
[ 3017.522272] FAT-fs (sdb2): invalid media value (0x0d)
[ 3017.522280] FAT-fs (sdb2): Can't find a valid FAT filesystem
[ 3017.536361] XFS (sdb2): Invalid superblock magic number
[ 3017.553342] FAT-fs (sdb2): utf8 is not a recommended IO charset for FAT filesystems, filesystem will be case sensitive!
[ 3017.564840] FAT-fs (sdb2): invalid media value (0x0d)
[ 3017.564852] FAT-fs (sdb2): Can't find a valid FAT filesystem
[ 3017.580862] MINIX-fs: unable to read superblock
[ 3017.586494] attempt to access beyond end of device
[ 3017.586508] sdb2: rw=16, want=3, limit=2
[ 3017.586516] hfsplus: unable to find HFS+ superblock
[ 3017.616705] qnx4: no qnx4 filesystem (no root dir).
[ 3017.622247] You didn't specify the type of your ufs filesystem

mount -t ufs -o ufstype=sun|sunx86|44bsd|ufs2|5xbsd|old|hp|nextstep|nextstep-cd|openstep ...

>>>WARNING<<< Wrong ufstype may corrupt your filesystem, default is ufstype=old
[ 3017.632980] hfs: can't find a HFS filesystem on dev sdb2
[ 4450.793736] perf interrupt took too long (2503 > 2500), lowering kernel.perf_event_max_sample_rate to 50000

The problem is... I never had that device before

Device Boot Start End Sectors Size Id Type
/dev/sdb1 * 2048 74219519 74217472 35,4G 83 Linux
/dev/sdb2 74221566 78241791 4020226 1,9G 5 Extended
/dev/sdb5 74221568 78241791 4020224 1,9G 82 Linux swap / Solaris
Looking at fdisk seems like it is "overlapping" the swap partition... is this a good thing ?
EDIT: now I wonder if the Grub problem was because of this new-born partition...

 

[levi]

After a reboot ssh is fine, also the timer magically appeared in Grub


Now I have shitty watered colors and a vibrating display... I almost can't read what I'm writing... let's return to the video drivers lectures...

Now you have ssh incoming working, you can fix the video drivers at your leisure from another box.

 

[FBnil]

"overlapping" the swap partition.

That is normal

old rant here:

Spoiler

Nope not good, fortunately:


1. sudo swapoff -a

2. Gparted, delete swap, create swap (not sure, but it needs a reboot to read the new partition...)


3. sudo swapon -a

Probably wont give you problems for now, just stay outside the 90% full.... but fix it soon...


Ah, did not read your post well. The /dev/sdb2 is an extended partition, and inside that, reside multiple partitions, in your case only one. The size is ok if you have 2Gb RAM. It is used to swap out big programs to Harddisk and to suspend to disk (sleep a machine).


Usually you create a user data partition and a separate root partition (for the OS), this minimizes time of "your machine has rebooted 31 times, a fsck is forced"... and then you have to wait a gazillion minutes for it to complete, if the OS is static, compartimentalization also solves a bit of the fragmentation you get, where OS is intertwined with your data,


Now, there is nothing bad with having /tmp/ and /home and /usr and /var in the same filesystem, but if yours gets full, downloading stuff... the system freezes.. because the OS requires space too.


However, it seems like your systems is trying to use those unexisting partitions...


Here is some background info: http://askubuntu.com/questions/151968/what-does-the-term-extended-partition-mean-is-it-safe-to-use-this-type-of-par


As for the errors:


use this command: lsblk   a normal partitionlooks like:  http://www.cyberciti.biz/faq/linux-list-disk-partitions-command/


Mine looks like:


sda                         8:0    0   1.8T  0 disk  


├─sda1                      8:1    0   243M  0 part  /boot


├─sda2                      8:2    0     1K  0 part  


└─sda5                      8:5    0   1.8T  0 part  


  └─sda5_crypt (dm-0)     254:0    0   1.8T  0 crypt


    ├─hostname-root (dm-1)   254:1    0   332M  0 lvm   /


    ├─hostname-usr (dm-2)    254:2    0   8.4G  0 lvm   /usr


    ├─hostname-var (dm-3)    254:3    0   2.8G  0 lvm   /var


    ├─hostname-swap_1 (dm-4) 254:4    0  15.7G  0 lvm   [sWAP]


    ├─hostname-tmp (dm-5)    254:5    0   380M  0 lvm   /tmp


    └─hostname-home (dm-6)   254:6    0   1.8T  0 lvm   /home

You can not mount swap partitions with the mount -t ufs command. Swapon is used.


for completion, here is my fdisk -l /dev/sda  very similar to yours...


   Device Boot      Start         End      Blocks   Id  System


/dev/sda1   *        2048      499711      248832   83  Linux


/dev/sda2          501758  3907028991  1953263617    5  Extended


/dev/sda5          501760  3907028991  1953263616   83  Linux


As you can see, It is Linux (type 83) not Linux Swap (type 82) as you have. A Linux system survives quite well without swap, just look at your OpenPandora.


Maybe you want to execute the command "swapoff -a", delete the swap partition and create a generic linux partition from it?


Now, that Linux partition, yes, that can be formatted into "ufs", or ext4, or jfs, xfs, reiser4, hfs, fat32, etc... you need gparted for that. (or mkfs.* which is from the command line)

 

[PowerGod]

@FBnil: I just recreated and reformatted the swap partition, after that I just had to change the UUID in fstab to make it working again, but now it is in use and I have no more scary messages about that partition in dmesg Thank you

Also thanks for the detailed info, maybe I'll do a better partitioning when I will get a newer machine... for now I'm just lucky that this old junk is still working so well

 

[FBnil]

ah.. I re-edited my previous post. It is cleaner now.

 

[PowerGod]

You can not mount swap partitions with the mount -t ufs command. Swapon is used.

.

.

. 

/dev/sda5          501760  3907028991  1953263616   83  Linux

As you can see, It is Linux (type 82?) not Linux Swap (type 83?) as you have. A Linux system survives quite well without swap, just look at your OpenPandora.

Maybe you want to execute the command "swapoff -a", delete the swap partition and create a generic linux partition from it?

Now, that Linux partition, yes, that can be formatted into "ufs", or ext4, or jfs, xfs, reiser4, hfs, fat32, etc... you need gparted for that. (or mkfs.* which is from the command line)

Actually I never mounted myself the swap partition, I just created the partition for it during the Debian installation setup, then it did all by itself... I wasn't even aware how to use a swap partition, it was just working... the problems started only after the dist-upgrade

I used the only "swap" partition type I found listed in gparted, there is no code (82/83) in that interface, also it doesn't ask in what way to format it, there's just the command "format partition"... so I did that

Anyway I created it so large because this PC has only 512 Mb of RAM... it's actually like a Desktop Pandora but I open more programs here than with the Pandora

 

[FBnil]

Linux (type 82?) not Linux Swap (type 83?)

D'oh... swapped the numbers around... sorry. Swap is 82.

it's actually like a Desktop Pandora

Then you will need a keyboard like this: http://www.nintendoworldreport.com/media/2389/1/9342.jpg

Also, experts say that there is an ideal ratio beween RAM and swapsize. You could make your swap smaller, and add an extra partition for your data in there.

is still working so well

Yeah... dont write off those oldies yet...

 

[PowerGod]

Then you will need a keyboard like this: http://www.nintendoworldreport.com/media/2389/1/9342.jpg

LOL It's layout can be discussed forever in this forum

 

[PowerGod]

I found out and resolved many issues after the dist-upgrade, but there are still many that affect the system in a really bad way... for example 90% of the time I shut down the PC from LXDE, it doesn't turn OFF... I just hear it elaborate, all the icons and pointers disappear from the desktop, and then I hear the hard disk shutting down, BUT the PC is still ON with the empty wallpaper still visible... then I need to HARD TURNING IT OFF

After some research I found out what seems a Mozilla bug (that existed in 2013 and was resurrected some months ago) that happens just before the shutdown... now, I only use IceDove, but I close it manually before the shutdown... so the issue seems related to the fact that this Mozilla code start something and then doesn't close it...

BUT, seems also that this issue is present only on systems that are using this systemd all are shouting about...

this is the error of the Mozilla bug:

console-kit-daemon[1769]: (process:31434): GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed
 
---

By the way I found also many errors like this in the logs:

systemd-udevd[31441]: failed to execute '/lib/udev/socket/org/freedesktop/hal/udev_event' 'socket/org/freedesktop/hal/udev_event': No such file or directory But in this case turned out that in the upgrade process they removed the use of Hal and didn't uninstalled it automatically...

so with these commands all errors should go away:

sudo apt-get remove hal
sudo apt-get purge hal
sudo apt-get autoremove
EDIT:

After further investigation on the shutdown issue I found out that the system, while turned off by the graphic interface is reacting almost as an Halt command without shutdown... now I must find what is the exact command that is launched...

 

[PowerGod]

There is a difference between a

sudo shutdown nowissued from a graphic terminal (xterm) and from a plain Linux Terminal session O_O
If I launch it from an xterm in LXDE, it removes all icons and shuts down the hard disk, but doesn't turn off the PC, and the last thing that remains there is the wallpaper...

instead, if I, with the graphic session still open, switch to a Linux Terminal and launch the same command, the computer shuts off completely...

Where should I have to go to find out what is happening ?

 

[levi]

Try 'shutdown -h now'.  I don't know why it's halting without the -h from a terminal, but I always issue it with the -h myself.

 

[PowerGod]

Try 'shutdown -h now'.  I don't know why it's halting without the -h from a terminal, but I always issue it with the -h myself.

Same behaviour... -_- and it's also the same as when I choose to shutdown from the LXDE menu

I searched in the environmental variables if there was something weird, but I just found a duplicated list of paths, maybe because of the distro upgrade

 

[TrashyMG]

have you tried 'halt -p' ?  

I know in linux it's just an alias for the shutdown command, but who knows it may behave differently.

 

[Björn]

There is a difference between a

sudo shutdown nowissued from a graphic terminal (xterm) and from a plain Linux Terminal session O_O

If I launch it from an xterm in LXDE, it removes all icons and shuts down the hard disk, but doesn't turn off the PC, and the last thing that remains there is the wallpaper...


instead, if I, with the graphic session still open, switch to a Linux Terminal and launch the same command, the computer shuts off completely...


Where should I have to go to find out what is happening ?

That sounds like something is wrong with your graphics driver. The reason why it shows the wallpaper is that the graphics driver probably breaks when Xorg is shutting down and then doesn't switch the mode back and probably just freezes.
Which graphics card and driver are you using?

 

[PowerGod]

have you tried 'halt -p' ?  

I know in linux it's just an alias for the shutdown command, but who knows it may behave differently.

Tried, same result

There is a difference between a

sudo shutdown nowissued from a graphic terminal (xterm) and from a plain Linux Terminal session O_O
If I launch it from an xterm in LXDE, it removes all icons and shuts down the hard disk, but doesn't turn off the PC, and the last thing that remains there is the wallpaper...

instead, if I, with the graphic session still open, switch to a Linux Terminal and launch the same command, the computer shuts off completely...

Where should I have to go to find out what is happening ?

That sounds like something is wrong with your graphics driver. The reason why it shows the wallpaper is that the graphics driver probably breaks when Xorg is shutting down and then doesn't switch the mode back and probably just freezes.Which graphics card and driver are you using?

I have an integrated NVidia GeForce 2 fx, and I suppose I'm using "nouveau"... I don't have an xorg.conf file, anyway even 3D acceleration works