Puppy Linux Discussion Forum

[puppyluvr]

Hello,
Today I received an error at boot concerning a deleted inode..
It happens.. Failure to unmount etc... No biggie.. just fsck it, right??
OK, so e2fsck reports the partition "Maybe non-existent" ..
Since it is ext3, I tried an ext2 partition.. It "Maybe non-existent" as well..
Bear in mind GParted, PMount, Rox, Thunar etc all see them fine..

I was Woofing a Pup with the 3.3.2 Kernel today, and testing the ISO`s as I went.. Until the inode error flat refused to allow booting unless it was either:
A CD/DVD/USB code "pfix=ram" and "nocopy"...
Any attempt to access the drive during boot produces a kernel panic..
Error says it forced an unmount to all partitions..
UNLESS it is an already existing savefile!!!

Can a kernel do permanent damage being booted frugal within a partition.. It is a Pemasu kernel, whom amongst 4 (Also BK, Wux and yes, l0wt3ch), build kernels I trust explicitly...
Should I test via CD? This has never been a problem before.. It is a 215gb partition that has the failed inodes...

So, like, it doesnt effect my day to day, but testing anything new means nocopy, no access... With a toram nocopy boot, the partitions are active, as with a frugal savefile boot, but I cannot fsck the error..

I dont have an external ATM to back up stuff, so if it refuses access altogether... I am FSCK`ed... 150+ gb of all things Puppy, gone...
A LOT of work.. A lot of useful, hard to replace stuff... So...

How can a partition exist to anything that already "exists" but not to anything new..
Why can I boot to an existing savefile, or with toram/nocopy, but not with anything new accessing the drive???

HELP...

[Karl Godt]

There might be problems with the fsck binaries or with the kernel driver .

Probably the kernel driver does not support some features of the fs (ie mount options) .

Maybe the fsck/e2fsk binary is buggy .

WHERE DO YOU WOOF FROM ?

Ubuntu, Slackware,local repo ??

Remember everything is in development . Though there had been not many real problems, new code might be too new .

I suspect some bugs in woof , too, or the user .
It seems that some kernel drivers are missing too in your initrd.gz .

I am running full installations only on my main machine and have only a 40GB storage partition on ext2 . ext3+4 have not given me many troubles .

*

[puppyluvr]

Hello,
Well, the problem exists in whatever Puppy I try to e2fsck it..
I was woofing from Lucid, with local packages, from which I have built several Pups, without trouble..
So it cant be fsck or the kernel, it has to be the partition itself..
Tried live CD boots, frugal boots, usb boots..
From within Lucid, Wary, Racy, Slacko and even 5.11...
Still I cant fsck that partition..
Your code produced about a dozen backup superblocks, which I am trying now.. Be back...

[aarf]

there is a "check" in gparted that accomplishes the task to a better degree.
just checking #fcsk -f /dev/sdb1 is the correct format

[bigpup]

Do these checks from a Puppy live CD using boot option puppy pfix=ram.
This makes sure the drive is not mounted.

There is a chance the hard drive is going bad.

[puppyluvr]

Hello,
I always fsck pfix=ram..
GParted removed the error on bootup... (Thank you aarf.. )
However e2fsck still returns:

[trapster]

[Karl Godt]

/dev/
missing

*

"e2fsck sda" would work if you lauch rxvt in the /dev/ directory :
bash-3.00# pwd
/dev
bash-3.00# e2fsck sda10
e2fsck 1.41.9 (22-Aug-2009)
sda10 has been mounted 37 times without being checked, check forced.


*

pdev1=sda3 is a bootparameter that is shorthanded for pdev1=/dev/sda3 for convenience .

Someone only booting frugal would probably never see the bootparameter root=/dev/sda3 which is common for full installations .

[puppyluvr]

Hello,
Duh, now how did I miss that..
5 years of Linux, and I forgot the /dev.....
Just goes to show, sometimes the simple things....