| Author |
Message |
ruebe99
|
|
Post subject: memtes+ makes system unbootable
 Posted: 01.11.2014, 19:52
|
|
Joined: 2010-09-11
Posts: 44
Status: Offline
|
|
grub / grub-mkconfig is generating thousands (!) of menue items for memtest84+ in grub.cfg which makes the system unbootable from harddisks. chmod -x for the 20_memtest84+ or even renaming it (xxx_20...) does not solve. no idea how to proceed.
cheers,
ruebe99 |
|
|
| |
|
|
|
 |
|
bfree
|
|
|
Post subject: RE: memtes+ makes system unbootable
Posted: 01.11.2014, 23:46
|
|
Team Member
Joined: 2010-08-26
Posts: 267
Status: Offline
|
|
Run update-grub after the chmod (or just remove memtest86+).
Of course you shouldn't be ending up with "thousands" of memtest86+ entries in grub so I wonder if you have suffered filesystem corruption or something :-/ You could try running "debsums -as" and see if it complains about a lot of stuff, particularly anything grub or memtest related. |
|
|
| |
|
|
|
|
|
ruebe99
|
|
|
Post subject: RE: memtes+ makes system unbootable
Posted: 02.11.2014, 20:47
|
|
Joined: 2010-09-11
Posts: 44
Status: Offline
|
|
I checked 2 disk: all partitions are fine.
actually the grub.cfg has 251583 lines with
grep memtest grub.cfg | wc -l resulting in 45212 hits - I assume half of them are menue entries.
Even with no 20_memtest in grub.d grub is still generating memtest statements in grub.cfg. Is that correct? |
|
|
| |
|
|
|
|
|
bfree
|
|
|
Post subject: Re: RE: memtes+ makes system unbootable
Posted: 03.11.2014, 03:03
|
|
Team Member
Joined: 2010-08-26
Posts: 267
Status: Offline
|
|
|
ruebe99 wrote:
Even with no 20_memtest in grub.d grub is still generating memtest statements in grub.cfg. Is that correct?
No, that is wrong assuming you ran update-grub after removing it. grub.cfg is made up by running the files in /etc/grub.d (and flags each section with the file which produced it, so something like:
### BEGIN /etc/grub.d/20_memtest86+ ###
blah blah
### END /etc/grub.d/20_memtest86+ ###
Should appear around the memtest entries to tell you where they came from.
You also didn't mention if you tried debsums to check if you have all sorts of corruption (it will list some things so I'll happily admit the command I gave means you need to interpret the results a bit to figure out if you really have a problem, just showing here what it mentions about grub though would probably be enough in this case). |
|
|
| |
|
|
|
|
|
ruebe99
|
|
|
Post subject: RE: Re: RE: memtes+ makes system unbootable
Posted: 03.11.2014, 11:44
|
|
Joined: 2010-09-11
Posts: 44
Status: Offline
|
|
sry - here the debsums:
debsums -as
debsums: missing file /etc/apache2/mods-available/ident.load (from apache2 package)
debsums: changed file /etc/modprobe.d/dkms.conf (from dkms package)
debsums: changed file /etc/default/dmakms (from dmakms package)
debsums: changed file /etc/sane.d/dll.conf (from libsane-common package)
for the grub.cfg here a snapshop. have a look at the line numbers.
251542 menuentry 'Debian GNU/Linux, kernel memtest86+ (auf /dev/sdb1) (auf /dev/sdi6) (auf /d ev/sdb3) (auf /dev/sdb1) (auf /dev/sdb3) (auf /dev/sdb1) (auf /dev/sdb3)' --class gnu-linux -- class gnu --class os $menuentry_id_option 'osprober-gnulinux-/boot/memtest86+.bin--e3145acd-04 01-4614-a854-9533a965805c' {
251543 insmod part_msdos
251544 insmod ext2
251545 set root='hd1,msdos3'
251546 if [ x$feature_platform_search_hint = xy ]; then
251547 search --no-floppy --fs-uuid --set=root --hint-bios=hd1,msdos3 --hint-efi=hd 1,msdos3 --hint-baremetal=ahci1,msdos3 e3145acd-0401-4614-a854-9533a965805c
251548 else
251549 search --no-floppy --fs-uuid --set=root e3145acd-0401-4614-a854-9533a965805c
251550 fi
251551 linux /boot/memtest86+.bin
251552 }
251553 menuentry 'Debian GNU/Linux, kernel memtest86+ (auf /dev/sdj1) (auf /dev/sdb3) (auf /d ev/sdb1) (auf /dev/sdb3) (auf /dev/sdb1) (auf /dev/sdb3)' --class gnu-linux --class gnu --clas s os $menuentry_id_option 'osprober-gnulinux-/boot/memtest86+.bin--e3145acd-0401-4614-a854-953 3a965805c' {
251554 insmod part_msdos
251555 insmod ext2 |
|
|
| |
|
|
|
|
|
bfree
|
|
|
Post subject: RE: Re: RE: memtes+ makes system unbootable
Posted: 03.11.2014, 14:03
|
|
Team Member
Joined: 2010-08-26
Posts: 267
Status: Offline
|
|
| The ### BEGIN and ### END lines which wrap the memtest entries are the really interesting bit though the above does reveal these lines aren't coming from 20_memtest86+ but from somewhere else (probably os-prober) as it is setting up /boot/memtest86+.bin to boot as if it was a Linux kernel. |
|
|
| |
|
|
|
|
|
ruebe99
|
|
|
Post subject: RE: Re: RE: memtes+ makes system unbootable
Posted: 03.11.2014, 20:45
|
|
Joined: 2010-09-11
Posts: 44
Status: Offline
|
|
Year - you are absolutely right! It's in the 30_ area
(251566 ### END /etc/grub.d/30_os-prober ###).
So moving the 20_memtest has really no effect.
Both disks (1 ssd) seem to be ok. all partitions passed without error.
One strange thing from syslog:
Nov 3 21:29:56 Cucumber10 udisksd[4067]: Error performing housekeeping for drive /org/freedesktop/UDisks2/drives/SAMSUNG_HM501II_S2HVJDQB303950: Error updating SMART data: sk_disk_check_sleep_mode: Operation not supported (udisks-error-quark, 0)
Nov 3 21:29:56 Cucumber10 udisksd[4067]: Error performing housekeeping for drive /org/freedesktop/UDisks2/drives/SAMSUNG_HD204UI_S2HGJ9JB902639: Error updating SMART data: sk_disk_check_sleep_mode: Operation not supported (udisks-error-quark, 0)
a
grep -i superblock is bringing up
Nov 2 18:54:21 Cucumber10 kernel: [ 211.762967] EXT4-fs (sda3): unable to read superblock
Nov 2 18:54:21 Cucumber10 kernel: [ 211.766521] EXT2-fs (sda3): error: unable to read superblock
Nov 2 18:54:21 Cucumber10 kernel: [ 211.767614] EXT3-fs (sda3): error: unable to read superblock
Nov 2 18:54:21 Cucumber10 kernel: [ 211.769014] XFS (sda3): Invalid superblock magic number
Nov 2 18:54:21 Cucumber10 kernel: [ 211.777479] MINIX-fs: unable to read superblock
Nov 2 18:54:21 Cucumber10 kernel: [ 211.780307] hfsplus: unable to find HFS+ superblock
Nov 2 18:54:22 Cucumber10 kernel: [ 212.741979] EXT4-fs (sdb4): unable to read superblock
Nov 2 18:54:22 Cucumber10 kernel: [ 212.746381] EXT2-fs (sdb4): error: unable to read superblock
Nov 2 18:54:22 Cucumber10 kernel: [ 212.747739] EXT3-fs (sdb4): error: unable to read superblock
Nov 2 18:54:22 Cucumber10 kernel: [ 212.749397] XFS (sdb4): Invalid superblock magic number
Nov 2 18:54:22 Cucumber10 kernel: [ 212.758223] MINIX-fs: unable to read superblock
Nov 2 18:54:22 Cucumber10 kernel: [ 212.759608] hfsplus: unable to find HFS+ superblock
Nov 2 20:35:47 Cucumber10 kernel: [ 2240.820414] EXT4-fs (sda3): unable to read superblock
Nov 2 20:35:47 Cucumber10 kernel: [ 2240.829218] EXT2-fs (sda3): error: unable to read superblock
Nov 2 20:35:47 Cucumber10 kernel: [ 2240.831666] EXT3-fs (sda3): error: unable to read superblock
Nov 2 20:35:47 Cucumber10 kernel: [ 2240.836922] XFS (sda3): Invalid superblock magic number
Nov 2 20:35:47 Cucumber10 kernel: [ 2240.847726] MINIX-fs: unable to read superblock
Nov 2 20:35:47 Cucumber10 kernel: [ 2240.851336] hfsplus: unable to find HFS+ superblock
Nov 2 20:35:48 Cucumber10 kernel: [ 2241.975372] EXT4-fs (sdb4): unable to read superblock
Nov 2 20:35:48 Cucumber10 kernel: [ 2241.979983] EXT2-fs (sdb4): error: unable to read superblock
Nov 2 20:35:48 Cucumber10 kernel: [ 2241.981672] EXT3-fs (sdb4): error: unable to read superblock
Nov 2 20:35:48 Cucumber10 kernel: [ 2241.983746] XFS (sdb4): Invalid superblock magic number
Nov 2 20:35:48 Cucumber10 kernel: [ 2241.995886] MINIX-fs: unable to read superblock
Nov 2 20:35:48 Cucumber10 kernel: [ 2241.997460] hfsplus: unable to find HFS+ superblock
in syslog.1
syslog has none of this entries.
fyi:
infobash -v3
Host/Kernel/OS "Cucumber10" running Linux 3.17-2.slh.2-aptosid-amd64 x86_64 [ aptosid 2011-03 Πόνος - kde-lite - (201112312100) ]
CPU Info 4x Intel Core2 Quad Q9550 @ 6144 KB cache flags( sse3 ht nx lm vmx ) clocked at [ 2000.000 MHz ]
Videocard Intel 4 Series Chipset Integrated Graphics Controller tty resolution ( 138x24 )
Network cards Intel 82567LM-3 Gigabit
Processes 319 | Uptime 1day | Memory 3094.8/7829.0MB | HDD ST31000528AS,SanDisk SDSSDX12,Transcend,3200BEV External,External USB 3.0,HD204UI,Basics Desktop,Basics Desktop Size 7465GB (66%used) | Runlevel 5 | Client Shell | Infobash v3.47 |
|
|
| |
|
|
|
|
|
ruebe99
|
|
|
Post subject: RE: Re: RE: memtes+ makes system unbootable
Posted: 15.11.2014, 23:53
|
|
Joined: 2010-09-11
Posts: 44
Status: Offline
|
|
done so far. deleted some partions with gparted and re-installed a few.
thanks for support.
most of system is operational.
no hardware prob has been identified. |
|
|
| |
|
|
|
|
|
heidieboe
|
|
|
Post subject: RE: Re: RE: memtes+ makes system unbootable
Posted: 03.12.2015, 17:32
|
|
Joined: 2011-01-09
Posts: 11
Location: Verden (Aller)
Status: Offline
|
|
the other day I ran into the same problem.
/boot/grub/grub.cfg grew to a size of 2.6 MB and rendered the laptop unbootable.
My setup:
A laptop with two harddisks, each containing two installations of aptosid, 32-bit and 64-bit versions on each of the disk. They all share a common /home on the second disk:
sda1: 32bit /
sda3: 64bit /
sdb1: 32bit /
sdb2: 64bit /
sdb3: /home
I DU the four installations regularly. The installationss on the second disk serve as backups, just in case something goes desperately wrong with dist-upgrades of the sda systems
Obviously, with each new aptosid kernel the grub.cfg files started to grow, with lots of lines contaning the
"(auf /dev/sdb1) (auf /dev/sdi6) (auf /dev/sdb3) (auf /dev/sdb1) (auf /dev/sdb3) (auf /dev/sdb1) (auf /dev/sdb3)"
type of pattern seen by ruebe99.
To solve the problem I added a line
GRUB_DISABLE_OS_PROBER=true
to each of the installations on (in my case) sda3, sdb1 and sdb2, did an "update-grub" with each of them and a final update-grub with the (unmodified /etc/default/grub) sda1 system.
To me it looks like update-grub generates myriads of menu items when other disks/partitions also contain grub.cfg files that refer to each other. Looks like a combinatorial explosion.
A fine tuned solution would probably involve a few GRUB_OS_PROBER_SKIP_LIST controls in the /etc/default/grub files involved.
It is maybe not a bug of update-grub but a feature that should generate a warning message, nonetheless.
dieter b. |
|
|
| |
|
|
|
|
|
|
