Puppy Linux Discussion Forum

[rhadon]

Hi,

inspired by some older threads from the beginning of 2010, I formatted one of my partitions with ReiserFS, installed Lupu-525 frugal and use it since some days without any problem.

I'm wondering that this is no standard option in Puppy and no one or only a few one are using it. It seems to be easy to implement. Are there any disadvantages I don't see?

Next part is using a savefile with ReiserFS.

Is there anybody who got it to work in a newer Puppy? With
convert-pupsave.tar.gz
from PaulBx1 (btw. a very useful program, Thanks ) I converted one savefile to .rfs. I changed init so I can see it in the list of savefiles and I'm able to choose it. But every attempt ends with Kernel Panic.
The code in init script seems to be too sophisticated for me. Any idea?

Also here the same question as above. Are there any disadvantages I don't see?

Thanks in advance

Rolf

[Karl Godt]

How many save-files do you have inside the folder and partition and whole disk ?

And what are their names ?

If the list is very long , attach it .

To me it looks like the save-file name does not get passed correctly

and instead some other save-file from the

FND_PUPSAVE2FS="`find /mnt/data -maxdepth 2 -xdev -type f -name ${DISTRO_FILE_PREFIX}save*.[234]fs | grep -v ' ' | sed -e 's%^/mnt/data%%g' | tr '\n' ' '`"

which you must have modified I guess to include .rfs .
Unfortunately I can not help if you dont attach your init .

[Béèm]

rhadon, it's good that people pioneer like you are doing.
In the create save file dialogue nor reiserfs nor etx4 are proposed.
This is because the developers haven't tested it thoroughly and aren't confident about the use.

I remember discussions about reiserfs in earlier threads.
I don't remember the reason(s) behind it, but I think reiserfs wasn't recommended.
Maybe you can find those threads through the puppy search link in my sig.

[Karl Godt]

And another thing to consider :

The Puppy initrd.gz uses a very old statically compiled busybox ( v .1.4 , yes 1.4 , not 1.14 ) .

In todays busybox make menuconfig there is

Symbol: MKFS_REISER
Symbol: FEATURE_VOLUMEID_REISERFS

I actually do not know if mount uses some binaries of the

reiserfsprogs-3.6.21

as mount helper .

"mount -i" is needed to ignore mount helpers like ntfs-3g .

I had once installed suse-9x on reiserfs 3.5 I think which came with kernel 2.6.9 .
Since it had no drivers for my modem and printer I abandoned it .

I actually don't know since which version the kernel supports reiserfs 3.6 .
I guess it is the same like squashfs version 3 until maybe kernel 2.6.29 and later squashfs 4 .

There is no menu point in the kernel make menuconfig to choose between versions of filesystems .
It is only squashfs=n|y|m or reiserfs=n|y|m .

And if compiled as module it has to be inside the initrd.gz .

Reiserfs support is compiled into the kernel of 2.6.30.5 and 2.6.33.2 .

But if you use a mkfs.reisefs version 3.5 to create the fs , you may get problems .

[Karl Godt]

My busybox -v 1.18.3 really does it on luci218/511 :

creating a save-file of approx 100MB :
# touch /reiser3.6.rfs
# dd if=/dev/zero of=/reiser3.6.rfs bs=512 count=200000

Making the file system works only with the '--force' option :
[ /sbin/mkfs.reiser: symbolic link to `../bin/busybox' ]
# mkfs.reiser -f /reiser3.6.rfs

Mounting the reiser-save-file :
# for i in `seq 0 15` ; do
[ ! -d /dev/loop ] && mkdir /dev/loop;
ln -s ../loop$i /dev/loop/$i;
done #this part I don't know why busybox wants /dev/loop/0 .. its like fbset

# mkdir /REIS
# losetup /dev/loop0 /reiser3.6.rfs
# busybox mount -o loop -t reiserfs /reiser3.6.rfs /REIS

mount shows :
/dev/loop/0 on /REIS type reiserfs (rw,relatime)

df -m :
/dev/loop/0 98 33 66 34% /REIS

So the journal seems to occupy a lot of space ...

And I am not sure about the reiserfs --version busybox creates .

[rhadon]

Thanks for the responds.

At last I've found my answer about disadvantages:

[Béèm]

rhadon, for the sake of fellow posters, could you edit the subject of the initial post and add [SOLVED] or [ABANDONED] at the end.